Merge branch 'main' into nightly

.github/ISSUE_TEMPLATE/bug_report.yml

@@ -49,7 +49,7 @@ body:
     attributes:
       label: nanobot Version
       description: Run `nanobot --version` or `pip show nanobot-ai`
-      placeholder: e.g., 0.1.5
+      placeholder: e.g., 0.2.0
     validations:
       required: true
 

Dockerfile

@@ -14,8 +14,9 @@ RUN apt-get update && \
 
 WORKDIR /app
 
-# Install Python dependencies first (cached layer)
-COPY pyproject.toml README.md LICENSE ./
+# Install Python dependencies first (cached layer). Hatch reads the custom build
+# hook from hatch_build.py even for this metadata-only install.
+COPY pyproject.toml README.md LICENSE THIRD_PARTY_NOTICES.md hatch_build.py ./
 RUN mkdir -p nanobot bridge && touch nanobot/__init__.py && \
     uv pip install --system --no-cache . && \
     rm -rf nanobot bridge
@@ -23,6 +24,7 @@ RUN mkdir -p nanobot bridge && touch nanobot/__init__.py && \
 # Copy the full source and install
 COPY nanobot/ nanobot/
 COPY bridge/ bridge/
+COPY webui/ webui/
 RUN uv pip install --system --no-cache .
 
 # Build the WhatsApp bridge
@@ -43,8 +45,8 @@ RUN sed -i 's/\r$//' /usr/local/bin/entrypoint.sh && chmod +x /usr/local/bin/ent
 USER nanobot
 ENV HOME=/home/nanobot
 
-# Gateway default port
-EXPOSE 18790
+# Gateway health endpoint and optional WebUI/WebSocket channel ports
+EXPOSE 18790 8765
 
 ENTRYPOINT ["entrypoint.sh"]
 CMD ["status"]

README.md

@@ -23,6 +23,7 @@
 
 ## 📢 News
 
+- **2026-05-15** 🚀 Released **v0.2.0** — **`/goal`** holds sustained objectives across turns, WebUI now ships inside the wheel, image generation end to end, 5 new providers with `fallback_models`, and a real agent-loop refactor. Please see [release notes](https://github.com/HKUDS/nanobot/releases/tag/v0.2.0) for details.
 - **2026-05-14** 🎯 **`/goal`** for long-term objectives, visible multi-step progress, long-horizon missions in chat.
 - **2026-05-13** 🧠 Streaming reasoning before answers, automatic backup models, smoother plug-in reconnects.
 - **2026-05-12** 🎛️ Saved model presets with WebUI badge, simpler plug-in tools, quieter Feishu topic threads.
@@ -214,10 +215,9 @@ nanobot agent
 - Want to run nanobot in chat apps like Telegram, Discord, WeChat or Feishu? See [Chat Apps](./docs/chat-apps.md)
 - Want Docker or Linux service deployment? See [Deployment](./docs/deployment.md)
 
-## 🧪 WebUI (Development)
+## 🌐 WebUI
 
-> [!NOTE]
-> The WebUI development workflow currently requires a source checkout and is not yet shipped together with the official packaged release. See [WebUI Document](./webui/README.md) for full WebUI development docs and build steps.
+The WebUI ships **inside the published wheel** — no extra build step. Just enable the WebSocket channel and open it in your browser.
 
 <p align="center">
   <img src="images/nanobot_webui.png" alt="nanobot webui preview" width="900">
@@ -235,13 +235,12 @@ nanobot agent
 nanobot gateway
 ```
 
-**3. Start the webui dev server**
+**3. Open the WebUI**
 
-```bash
-cd webui
-bun install
-bun run dev
-```
+Visit [`http://127.0.0.1:8765`](http://127.0.0.1:8765) in your browser. To open it from another device on your LAN, see [WebUI docs → LAN access](./webui/README.md#access-from-another-device-lan).
+
+> [!TIP]
+> Working on the WebUI itself? Check out [`webui/README.md`](./webui/README.md) for the Vite dev server (HMR) workflow.
 
 ## 🏗️ Architecture
 

docker-compose.yml

@@ -20,6 +20,7 @@ services:
     restart: unless-stopped
     ports:
       - 18790:18790
+      - 8765:8765
     deploy:
       resources:
         limits:

docs/README.md

@@ -15,6 +15,7 @@ Start here for setup, everyday usage, and deployment.
 | Agent social network | [`agent-social-network.md`](./agent-social-network.md) | Join external agent communities from nanobot |
 | Configuration | [`configuration.md`](./configuration.md) | Providers, tools, channels, MCP, and runtime settings |
 | Image generation | [`image-generation.md`](./image-generation.md) | Configure image providers, WebUI image mode, and generated artifacts |
+| WebUI | [`../webui/README.md`](../webui/README.md) | Open the bundled browser UI; LAN access; Vite dev server for contributors |
 | Multiple instances | [`multiple-instances.md`](./multiple-instances.md) | Run isolated bots with separate configs and workspaces |
 | CLI reference | [`cli-reference.md`](./cli-reference.md) | Core CLI commands and common entrypoints |
 | In-chat commands | [`chat-commands.md`](./chat-commands.md) | Slash commands and periodic task behavior |

docs/configuration.md

@@ -26,7 +26,52 @@ Instead of storing secrets directly in `config.json`, you can use `${VAR_NAME}`
 }
 ```
 
-For **systemd** deployments, use `EnvironmentFile=` in the service unit to load variables from a file that only the deploying user can read:
+Any string value in `config.json` can use `${VAR_NAME}`. Resolution runs once at startup, in memory only — resolved values are never written back to disk, so editing config through `nanobot onboard` or the WebUI preserves the placeholder.
+
+If a referenced variable is unset, nanobot fails fast at startup with `ValueError: Environment variable 'NAME' referenced in config is not set`.
+
+### More examples
+
+**MCP servers** — both stdio `env` and HTTP `headers`:
+
+```json
+{
+  "tools": {
+    "mcpServers": {
+      "github": {
+        "command": "npx",
+        "args": ["-y", "@modelcontextprotocol/server-github"],
+        "env": { "GITHUB_PERSONAL_ACCESS_TOKEN": "${GITHUB_TOKEN}" }
+      },
+      "remote": {
+        "url": "https://example.com/mcp/",
+        "headers": { "Authorization": "Bearer ${REMOTE_MCP_TOKEN}" }
+      }
+    }
+  }
+}
+```
+
+**Web search providers:**
+
+```json
+{
+  "tools": {
+    "web": {
+      "search": {
+        "provider": "brave",
+        "apiKey": "${BRAVE_API_KEY}"
+      }
+    }
+  }
+}
+```
+
+### Loading variables at startup
+
+Pick whatever fits your deployment — nanobot only reads `os.environ` at startup, so any mechanism that populates the process environment works.
+
+**systemd** — use `EnvironmentFile=` in the service unit to load variables from a file that only the deploying user can read:
 
 ```ini
 # /etc/systemd/system/nanobot.service (excerpt)
@@ -42,6 +87,35 @@ TELEGRAM_TOKEN=your-token-here
 IMAP_PASSWORD=your-password-here
 ```
 
+**Docker** — pass an env file to the locally built image (one `KEY=VALUE` per line), or use `-e KEY=value`:
+
+```bash
+docker run --rm --env-file=./nanobot.env \
+  -v ~/.nanobot:/home/nanobot/.nanobot \
+  nanobot agent -m "Hello"
+```
+
+**direnv** — drop a `.envrc` in your working directory and run `direnv allow`:
+
+```bash
+# .envrc (auto-loaded by direnv)
+export TELEGRAM_TOKEN=your-token-here
+export ANTHROPIC_API_KEY=...
+```
+
+**Secret managers (1Password, Bitwarden, pass)** — wrap the process so secrets only exist as env vars for the lifetime of the run, never on disk:
+
+```bash
+# 1Password — references in .env.tpl look like `op://Vault/Item/field`
+op run --env-file=.env.tpl -- nanobot agent
+
+# pass (passwordstore.org)
+ANTHROPIC_API_KEY="$(pass show api/anthropic)" nanobot agent
+
+# Bitwarden
+ANTHROPIC_API_KEY="$(bw get password api/anthropic)" nanobot agent
+```
+
 ## Providers
 
 > [!TIP]
@@ -917,7 +991,7 @@ By default, web search uses `duckduckgo`, and it works out of the box without an
     "web": {
       "search": {
         "provider": "brave",
-        "apiKey": "BSA..."
+        "apiKey": "${BRAVE_API_KEY}"
       }
     }
   }
@@ -931,7 +1005,7 @@ By default, web search uses `duckduckgo`, and it works out of the box without an
     "web": {
       "search": {
         "provider": "tavily",
-        "apiKey": "tvly-..."
+        "apiKey": "${TAVILY_API_KEY}"
       }
     }
   }
@@ -945,7 +1019,7 @@ By default, web search uses `duckduckgo`, and it works out of the box without an
     "web": {
       "search": {
         "provider": "jina",
-        "apiKey": "jina_..."
+        "apiKey": "${JINA_API_KEY}"
       }
     }
   }
@@ -959,7 +1033,7 @@ By default, web search uses `duckduckgo`, and it works out of the box without an
     "web": {
       "search": {
         "provider": "kagi",
-        "apiKey": "your-kagi-api-key"
+        "apiKey": "${KAGI_API_KEY}"
       }
     }
   }
@@ -973,7 +1047,7 @@ By default, web search uses `duckduckgo`, and it works out of the box without an
     "web": {
       "search": {
         "provider": "olostep",
-        "apiKey": "YOUR_OLOSTEP_API_KEY"
+        "apiKey": "${OLOSTEP_API_KEY}"
       }
     }
   }
@@ -1136,6 +1210,8 @@ MCP tools are automatically discovered and registered on startup. The LLM can us
 > [!TIP]
 > For production deployments, set `"restrictToWorkspace": true` and `"tools.exec.sandbox": "bwrap"` in your config to sandbox the agent.
 
+For API keys, tokens, and other secrets, see [Environment Variables for Secrets](#environment-variables-for-secrets) — avoid storing them directly in `config.json`.
+
 | Option | Default | Description |
 |--------|---------|-------------|
 | `tools.restrictToWorkspace` | `false` | When `true`, restricts **all** agent tools (shell, file read/write/edit, list) to the workspace directory. Prevents path traversal and out-of-scope access. |

docs/deployment.md

@@ -10,6 +10,18 @@
 > [!IMPORTANT]
 > Official Docker usage currently means building from this repository with the included `Dockerfile`. Docker Hub images under third-party namespaces are not maintained or verified by HKUDS/nanobot; do not mount API keys or bot tokens into them unless you trust the publisher.
 
+> [!IMPORTANT]
+> The gateway and WebSocket channel default to `host: "127.0.0.1"` in `config.json` (set in `nanobot/config/schema.py`). Docker `-p` port forwarding cannot reach a container's loopback interface, so for the host or LAN to reach the exposed ports you must set both binds to `0.0.0.0` in `~/.nanobot/config.json` before starting the container:
+>
+> ```json
+> {
+>   "gateway":  { "host": "0.0.0.0" },
+>   "channels": { "websocket": { "host": "0.0.0.0" } }
+> }
+> ```
+>
+> When `host` is `0.0.0.0`, the gateway refuses to start unless `token` or `tokenIssueSecret` is also configured on the WebSocket channel — see [`webui/README.md`](../webui/README.md) for details.
+
 ### Docker Compose
 
 ```bash
@@ -36,8 +48,20 @@ docker run -v ~/.nanobot:/home/nanobot/.nanobot --rm nanobot onboard
 # Edit config on host to add API keys
 vim ~/.nanobot/config.json
 
-# Run gateway (connects to enabled channels, e.g. Telegram/Discord/Mochat)
-docker run -v ~/.nanobot:/home/nanobot/.nanobot -p 18790:18790 nanobot gateway
+# Run gateway (connects to enabled channels, e.g. Telegram/Discord/Mochat).
+# Mirrors the security caps and port mappings declared in docker-compose.yml:
+#   - `--cap-drop ALL --cap-add SYS_ADMIN` + unconfined apparmor/seccomp are required
+#     when `tools.exec.sandbox: "bwrap"` is enabled (bwrap needs CAP_SYS_ADMIN for
+#     user namespaces). Without them, `bwrap` exits with `clone3: Operation not permitted`.
+#   - `-p 8765:8765` exposes the WebSocket channel / WebUI alongside the gateway health
+#     endpoint on 18790.
+docker run \
+  --cap-drop ALL --cap-add SYS_ADMIN \
+  --security-opt apparmor=unconfined \
+  --security-opt seccomp=unconfined \
+  -v ~/.nanobot:/home/nanobot/.nanobot \
+  -p 18790:18790 -p 8765:8765 \
+  nanobot gateway
 
 # Or run a single command
 docker run -v ~/.nanobot:/home/nanobot/.nanobot --rm nanobot agent -m "Hello!"

hatch_build.py

@@ -0,0 +1,101 @@
+"""Hatch build hook that bundles the webui (Vite) into nanobot/web/dist.
+
+Triggered automatically by `python -m build` (and any other hatch-driven build)
+so published wheels and sdists ship a fresh webui without requiring developers
+to remember `cd webui && bun run build` beforehand.
+
+Behaviour:
+
+- Skips for editable installs (`pip install -e .`). Editable mode is for Python
+  development; webui contributors use `cd webui && bun run dev` (Vite HMR) and
+  do not need a packaged `dist/`.
+- No-op when `webui/package.json` is absent (e.g. installing from an sdist that
+  already contains a prebuilt `nanobot/web/dist/`).
+- Skips when `NANOBOT_SKIP_WEBUI_BUILD=1` is set.
+- Skips when `nanobot/web/dist/index.html` already exists, unless
+  `NANOBOT_FORCE_WEBUI_BUILD=1` is set.
+- Uses `bun` when available, otherwise falls back to `npm`. The chosen tool
+  performs `install` followed by `run build`.
+"""
+
+from __future__ import annotations
+
+import os
+import shutil
+import subprocess
+from pathlib import Path
+
+from hatchling.builders.hooks.plugin.interface import BuildHookInterface
+
+
+class WebUIBuildHook(BuildHookInterface):
+    PLUGIN_NAME = "webui-build"
+
+    def initialize(self, version: str, build_data: dict) -> None:  # noqa: D401
+        root = Path(self.root)
+        webui_dir = root / "webui"
+        package_json = webui_dir / "package.json"
+        dist_dir = root / "nanobot" / "web" / "dist"
+        index_html = dist_dir / "index.html"
+
+        # `pip install -e .` builds an editable wheel; skip the (slow) webui
+        # bundle since editable installs target Python development and webui
+        # work uses `bun run dev` instead.
+        if self.target_name == "wheel" and version == "editable":
+            self.app.display_info(
+                "[webui-build] skipped for editable install "
+                "(use `cd webui && bun run build` to bundle webui manually)"
+            )
+            return
+
+        if os.environ.get("NANOBOT_SKIP_WEBUI_BUILD") == "1":
+            self.app.display_info("[webui-build] skipped via NANOBOT_SKIP_WEBUI_BUILD=1")
+            return
+
+        if not package_json.is_file():
+            self.app.display_info(
+                "[webui-build] no webui/ source tree, assuming prebuilt nanobot/web/dist/"
+            )
+            return
+
+        force = os.environ.get("NANOBOT_FORCE_WEBUI_BUILD") == "1"
+        if index_html.is_file() and not force:
+            self.app.display_info(
+                f"[webui-build] reusing existing build at {dist_dir} "
+                "(set NANOBOT_FORCE_WEBUI_BUILD=1 to rebuild)"
+            )
+            return
+
+        runner = self._pick_runner()
+        if runner is None:
+            raise RuntimeError(
+                "[webui-build] neither `bun` nor `npm` is available on PATH; "
+                "install one or set NANOBOT_SKIP_WEBUI_BUILD=1 to bypass."
+            )
+
+        self.app.display_info(f"[webui-build] using {runner} to build webui")
+        self._run([runner, "install"], cwd=webui_dir)
+        self._run([runner, "run", "build"], cwd=webui_dir)
+
+        if not index_html.is_file():
+            raise RuntimeError(
+                f"[webui-build] build finished but {index_html} is missing; "
+                "check webui/vite.config.ts outDir."
+            )
+        self.app.display_info(f"[webui-build] webui ready at {dist_dir}")
+
+    @staticmethod
+    def _pick_runner() -> str | None:
+        for candidate in ("bun", "npm"):
+            if shutil.which(candidate):
+                return candidate
+        return None
+
+    def _run(self, cmd: list[str], *, cwd: Path) -> None:
+        self.app.display_info(f"[webui-build] $ {' '.join(cmd)} (cwd={cwd})")
+        try:
+            subprocess.run(cmd, cwd=cwd, check=True)
+        except subprocess.CalledProcessError as exc:
+            raise RuntimeError(
+                f"[webui-build] command failed ({exc.returncode}): {' '.join(cmd)}"
+            ) from exc

nanobot/__init__.py

@@ -21,7 +21,7 @@ def _resolve_version() -> str:
         return _pkg_version("nanobot-ai")
     except PackageNotFoundError:
         # Source checkouts often import nanobot without installed dist-info.
-        return _read_pyproject_version() or "0.1.5.post3"
+        return _read_pyproject_version() or "0.2.0"
 
 
 __version__ = _resolve_version()

nanobot/agent/autocompact.py

@@ -4,7 +4,7 @@ from __future__ import annotations
 
 from collections.abc import Collection
 from datetime import datetime
-from typing import TYPE_CHECKING, Any, Callable, Coroutine
+from typing import TYPE_CHECKING, Callable, Coroutine
 
 from loguru import logger
 
@@ -37,27 +37,6 @@ class AutoCompact:
     def _format_summary(text: str, last_active: datetime) -> str:
         return f"Previous conversation summary (last active {last_active.isoformat()}):\n{text}"
 
-    def _split_unconsolidated(
-        self, session: Session,
-    ) -> tuple[list[dict[str, Any]], list[dict[str, Any]]]:
-        """Split live session tail into archiveable prefix and retained recent suffix."""
-        tail = list(session.messages[session.last_consolidated:])
-        if not tail:
-            return [], []
-
-        probe = Session(
-            key=session.key,
-            messages=tail.copy(),
-            created_at=session.created_at,
-            updated_at=session.updated_at,
-            metadata={},
-            last_consolidated=0,
-        )
-        probe.retain_recent_legal_suffix(self._RECENT_SUFFIX_MESSAGES)
-        kept = probe.messages
-        cut = len(tail) - len(kept)
-        return tail[:cut], kept
-
     def check_expired(self, schedule_background: Callable[[Coroutine], None],
                       active_session_keys: Collection[str] = ()) -> None:
         """Schedule archival for idle sessions, skipping those with in-flight agent tasks."""
@@ -74,33 +53,17 @@ class AutoCompact:
 
     async def _archive(self, key: str) -> None:
         try:
-            self.sessions.invalidate(key)
-            session = self.sessions.get_or_create(key)
-            archive_msgs, kept_msgs = self._split_unconsolidated(session)
-            if not archive_msgs and not kept_msgs:
-                session.updated_at = datetime.now()
-                self.sessions.save(session)
-                return
-
-            last_active = session.updated_at
-            summary = ""
-            if archive_msgs:
-                summary = await self.consolidator.archive(archive_msgs) or ""
+            summary = await self.consolidator.compact_idle_session(
+                key, self._RECENT_SUFFIX_MESSAGES,
+            )
             if summary and summary != "(nothing)":
-                self._summaries[key] = (summary, last_active)
-                session.metadata["_last_summary"] = {"text": summary, "last_active": last_active.isoformat()}
-            session.messages = kept_msgs
-            session.last_consolidated = 0
-            session.updated_at = datetime.now()
-            self.sessions.save(session)
-            if archive_msgs:
-                logger.info(
-                    "Auto-compact: archived {} (archived={}, kept={}, summary={})",
-                    key,
-                    len(archive_msgs),
-                    len(kept_msgs),
-                    bool(summary),
-                )
+                session = self.sessions.get_or_create(key)
+                meta = session.metadata.get("_last_summary")
+                if isinstance(meta, dict):
+                    self._summaries[key] = (
+                        meta["text"],
+                        datetime.fromisoformat(meta["last_active"]),
+                    )
         except Exception:
             logger.exception("Auto-compact: failed for {}", key)
         finally:

nanobot/agent/loop.py

@@ -33,7 +33,6 @@ from nanobot.config.schema import AgentDefaults, ModelPresetConfig
 from nanobot.providers.base import LLMProvider
 from nanobot.providers.factory import ProviderSnapshot
 from nanobot.session.goal_state import (
-    goal_state_ws_blob,
     runner_wall_llm_timeout_s,
 )
 from nanobot.session.manager import Session, SessionManager
@@ -42,10 +41,14 @@ from nanobot.utils.document import extract_documents
 from nanobot.utils.helpers import image_placeholder_text
 from nanobot.utils.helpers import truncate_text as truncate_text_fn
 from nanobot.utils.image_generation_intent import image_generation_prompt
+from nanobot.utils.llm_runtime import LLMRuntime
 from nanobot.utils.runtime import EMPTY_FINAL_RESPONSE_MESSAGE
 from nanobot.utils.session_attachments import merge_turn_media_into_last_assistant
-from nanobot.utils.webui_titles import mark_webui_session, maybe_generate_webui_title_after_turn
-from nanobot.utils.webui_turn_helpers import publish_turn_run_status
+from nanobot.utils.webui_turn_helpers import (
+    WebuiTurnCoordinator,
+    build_bus_progress_callback,
+    mark_webui_session,
+)
 
 if TYPE_CHECKING:
     from nanobot.config.schema import (
@@ -136,6 +139,11 @@ class AgentLoop:
     def tool_names(self) -> list[str]:
         return self.tools.tool_names
 
+    def llm_runtime(self) -> LLMRuntime:
+        """Return the current provider/model pair owned by this loop."""
+        self._refresh_provider_snapshot()
+        return LLMRuntime(self.provider, self.model)
+
     _RUNTIME_CHECKPOINT_KEY = "runtime_checkpoint"
     _PENDING_USER_TURN_KEY = "pending_user_turn"
 
@@ -237,6 +245,11 @@ class AgentLoop:
 
         self.context = ContextBuilder(workspace, timezone=timezone, disabled_skills=disabled_skills)
         self.sessions = session_manager or SessionManager(workspace)
+        self._webui_turns = WebuiTurnCoordinator(
+            bus=self.bus,
+            sessions=self.sessions,
+            schedule_background=lambda coro: self._schedule_background(coro),
+        )
         self.tools = ToolRegistry()
         # One file-read/write tracker per logical session. The tool registry is
         # shared by this loop, so tools resolve the active state via contextvars.
@@ -524,34 +537,7 @@ class AgentLoop:
         self, msg: InboundMessage
     ) -> Callable[..., Awaitable[None]]:
         """Build a progress callback that publishes to the message bus."""
-
-        async def _bus_progress(
-            content: str,
-            *,
-            tool_hint: bool = False,
-            tool_events: list[dict[str, Any]] | None = None,
-            reasoning: bool = False,
-            reasoning_end: bool = False,
-        ) -> None:
-            meta = dict(msg.metadata or {})
-            meta["_progress"] = True
-            meta["_tool_hint"] = tool_hint
-            if reasoning:
-                meta["_reasoning_delta"] = True
-            if reasoning_end:
-                meta["_reasoning_end"] = True
-            if tool_events:
-                meta["_tool_events"] = tool_events
-            await self.bus.publish_outbound(
-                OutboundMessage(
-                    channel=msg.channel,
-                    chat_id=msg.chat_id,
-                    content=content,
-                    metadata=meta,
-                )
-            )
-
-        return _bus_progress
+        return build_bus_progress_callback(self.bus, msg)
 
     async def _build_retry_wait_callback(
         self, msg: InboundMessage
@@ -938,38 +924,12 @@ class AgentLoop:
                             content="", metadata=msg.metadata or {},
                         ))
                     if msg.channel == "websocket":
-                        # Signal that the turn is fully complete (all tools executed,
-                        # final text streamed).  This lets WS clients know when to
-                        # definitively stop the loading indicator.
                         turn_lat = self._pending_turn_latency_ms.pop(session_key, None)
-                        turn_metadata: dict[str, Any] = {**msg.metadata, "_turn_end": True}
-                        if turn_lat is not None:
-                            turn_metadata["latency_ms"] = int(turn_lat)
-                        sess_turn = self.sessions.get_or_create(session_key)
-                        turn_metadata["goal_state"] = goal_state_ws_blob(sess_turn.metadata)
-                        await self.bus.publish_outbound(OutboundMessage(
-                            channel=msg.channel, chat_id=msg.chat_id,
-                            content="", metadata=turn_metadata,
-                        ))
-                        if msg.metadata.get("webui") is True:
-                            async def _generate_title_and_notify() -> None:
-                                generated = await maybe_generate_webui_title_after_turn(
-                                    channel=msg.channel,
-                                    metadata=msg.metadata,
-                                    sessions=self.sessions,
-                                    session_key=session_key,
-                                    provider=self.provider,
-                                    model=self.model,
-                                )
-                                if generated:
-                                    await self.bus.publish_outbound(OutboundMessage(
-                                        channel=msg.channel,
-                                        chat_id=msg.chat_id,
-                                        content="",
-                                        metadata={**msg.metadata, "_session_updated": True},
-                                    ))
-
-                            self._schedule_background(_generate_title_and_notify())
+                        await self._webui_turns.handle_turn_end(
+                            msg,
+                            session_key=session_key,
+                            latency_ms=turn_lat,
+                        )
                 except asyncio.CancelledError:
                     logger.info("Task cancelled for session {}", session_key)
                     # Preserve partial context from the interrupted turn so
@@ -1021,8 +981,9 @@ class AgentLoop:
                         "Re-published {} leftover message(s) to bus for session {}",
                         leftover, session_key,
                     )
-            await publish_turn_run_status(self.bus, msg, "idle")
+            await self._webui_turns.publish_run_status(msg, "idle")
             self._pending_turn_latency_ms.pop(session_key, None)
+            self._webui_turns.discard(session_key)
 
     async def close_mcp(self) -> None:
         """Drain pending background archives, then close MCP connections."""
@@ -1338,6 +1299,11 @@ class AgentLoop:
             "include_timestamps": True,
         }
         ctx.history = ctx.session.get_history(**_hist_kwargs)
+        self._webui_turns.capture_title_context(
+            ctx.session_key,
+            ctx.msg,
+            self.llm_runtime(),
+        )
 
         ctx.initial_messages = self._build_initial_messages(
             ctx.msg, ctx.session, ctx.history, ctx.pending_summary
@@ -1354,7 +1320,7 @@ class AgentLoop:
         return "ok"
 
     async def _state_run(self, ctx: TurnContext) -> str:
-        await publish_turn_run_status(self.bus, ctx.msg, "running")
+        await self._webui_turns.publish_run_status(ctx.msg, "running")
         result = await self._run_agent_loop(
             ctx.initial_messages,
             on_progress=ctx.on_progress,

nanobot/agent/memory.py

@@ -678,11 +678,18 @@ class Consolidator:
         The budget reserves space for completion tokens and a safety buffer
         so the LLM request never exceeds the context window.
         """
-        if not session.messages or self.context_window_tokens <= 0:
+        if self.context_window_tokens <= 0:
             return
 
         lock = self.get_lock(session.key)
         async with lock:
+            # Refresh session reference: AutoCompact may have replaced it.
+            fresh = self.sessions.get_or_create(session.key)
+            if fresh is not session:
+                session = fresh
+            if not session.messages:
+                return
+
             budget = self._input_token_budget
             target = int(budget * self.consolidation_ratio)
             last_summary = await self._consolidate_replay_overflow(
@@ -769,6 +776,74 @@ class Consolidator:
             # the summary injection strategy with AutoCompact._archive().
             self._persist_last_summary(session, last_summary)
 
+    async def compact_idle_session(
+        self,
+        session_key: str,
+        max_suffix: int = 8,
+    ) -> str | None:
+        """Hard-truncate an idle session under the consolidation lock.
+
+        Used by AutoCompact so all session mutation goes through a single
+        lock-protected path.  Returns the summary text on success, ``None``
+        if the LLM failed (raw_archive fallback), or ``""`` if there was
+        nothing to archive.
+        """
+        lock = self.get_lock(session_key)
+        async with lock:
+            self.sessions.invalidate(session_key)
+            session = self.sessions.get_or_create(session_key)
+
+            tail = list(session.messages[session.last_consolidated:])
+            if not tail:
+                session.updated_at = datetime.now()
+                self.sessions.save(session)
+                return ""
+
+            probe = Session(
+                key=session.key,
+                messages=tail.copy(),
+                created_at=session.created_at,
+                updated_at=session.updated_at,
+                metadata={},
+                last_consolidated=0,
+            )
+            probe.retain_recent_legal_suffix(max_suffix)
+            kept = probe.messages
+            cut = len(tail) - len(kept)
+            archive_msgs = tail[:cut]
+
+            if not archive_msgs and not kept:
+                session.updated_at = datetime.now()
+                self.sessions.save(session)
+                return ""
+
+            last_active = session.updated_at
+            summary: str | None = ""
+            if archive_msgs:
+                summary = await self.archive(archive_msgs)
+
+            if summary and summary != "(nothing)":
+                session.metadata["_last_summary"] = {
+                    "text": summary,
+                    "last_active": last_active.isoformat(),
+                }
+
+            session.messages = kept
+            session.last_consolidated = 0
+            session.updated_at = datetime.now()
+            self.sessions.save(session)
+
+            if archive_msgs:
+                logger.info(
+                    "Idle-session compact for {}: archived={}, kept={}, summary={}",
+                    session_key,
+                    len(archive_msgs),
+                    len(kept),
+                    bool(summary),
+                )
+
+            return summary
+
 
 # ---------------------------------------------------------------------------
 # Dream — heavyweight cron-scheduled memory consolidation

nanobot/agent/runner.py

@@ -15,6 +15,12 @@ from loguru import logger
 from nanobot.agent.hook import AgentHook, AgentHookContext
 from nanobot.agent.tools.registry import ToolRegistry
 from nanobot.providers.base import LLMProvider, LLMResponse, ToolCallRequest
+from nanobot.utils.file_edit_events import (
+    build_file_edit_end_event,
+    build_file_edit_error_event,
+    build_file_edit_start_event,
+    prepare_file_edit_tracker,
+)
 from nanobot.utils.helpers import (
     IncrementalThinkExtractor,
     build_assistant_message,
@@ -26,6 +32,10 @@ from nanobot.utils.helpers import (
     strip_think,
     truncate_text,
 )
+from nanobot.utils.progress_events import (
+    invoke_file_edit_progress,
+    on_progress_accepts_file_edit_events,
+)
 from nanobot.utils.prompt_templates import render_template
 from nanobot.utils.runtime import (
     EMPTY_FINAL_RESPONSE_MESSAGE,
@@ -813,6 +823,30 @@ class AgentRunner:
             return prep_error + hint, event, (
                 RuntimeError(prep_error) if spec.fail_on_tool_error else None
             )
+        emit_file_edit_events = (
+            spec.progress_callback is not None
+            and on_progress_accepts_file_edit_events(spec.progress_callback)
+        )
+        progress_callback = spec.progress_callback if emit_file_edit_events else None
+        file_edit_tracker = (
+            prepare_file_edit_tracker(
+                call_id=tool_call.id,
+                tool_name=tool_call.name,
+                tool=tool,
+                workspace=spec.workspace,
+                params=params if isinstance(params, dict) else None,
+            )
+            if progress_callback is not None
+            else None
+        )
+        if file_edit_tracker is not None and progress_callback is not None:
+            await invoke_file_edit_progress(
+                progress_callback,
+                [build_file_edit_start_event(
+                    file_edit_tracker,
+                    params if isinstance(params, dict) else None,
+                )],
+            )
         try:
             if tool is not None:
                 result = await tool.execute(**params)
@@ -821,6 +855,11 @@ class AgentRunner:
         except asyncio.CancelledError:
             raise
         except BaseException as exc:
+            if file_edit_tracker is not None and progress_callback is not None:
+                await invoke_file_edit_progress(
+                    progress_callback,
+                    [build_file_edit_error_event(file_edit_tracker, str(exc))],
+                )
             event = {
                 "name": tool_call.name,
                 "status": "error",
@@ -842,6 +881,11 @@ class AgentRunner:
             return payload, event, None
 
         if isinstance(result, str) and result.startswith("Error"):
+            if file_edit_tracker is not None and progress_callback is not None:
+                await invoke_file_edit_progress(
+                    progress_callback,
+                    [build_file_edit_error_event(file_edit_tracker, result)],
+                )
             event = {
                 "name": tool_call.name,
                 "status": "error",
@@ -860,6 +904,12 @@ class AgentRunner:
                 return result + hint, event, RuntimeError(result)
             return result + hint, event, None
 
+        if file_edit_tracker is not None and progress_callback is not None:
+            await invoke_file_edit_progress(
+                progress_callback,
+                [build_file_edit_end_event(file_edit_tracker)],
+            )
+
         detail = "" if result is None else str(result)
         detail = detail.replace("\n", " ").strip()
         if not detail:

nanobot/channels/websocket.py

@@ -230,6 +230,25 @@ def _mask_secret_hint(secret: str | None) -> str | None:
     return f"{secret[:4]}••••{secret[-4:]}"
 
 
+def _provider_requires_api_key(spec: Any) -> bool:
+    if spec.backend == "azure_openai":
+        return True
+    if spec.is_local or spec.is_direct:
+        return False
+    return True
+
+
+def _provider_configured_for_settings(spec: Any, provider_config: Any) -> bool:
+    if _provider_requires_api_key(spec):
+        return bool(provider_config.api_key)
+    return bool(
+        provider_config.api_key
+        or provider_config.api_base
+        or getattr(provider_config, "region", None)
+        or getattr(provider_config, "profile", None)
+    )
+
+
 _WEB_SEARCH_PROVIDER_OPTIONS: tuple[dict[str, str], ...] = (
     {"name": "duckduckgo", "label": "DuckDuckGo", "credential": "none"},
     {"name": "brave", "label": "Brave Search", "credential": "api_key"},
@@ -786,13 +805,14 @@ class WebSocketChannel(BaseChannel):
         providers = []
         for spec in PROVIDERS:
             provider_config = getattr(config.providers, spec.name, None)
-            if provider_config is None or spec.is_oauth or spec.is_local:
+            if provider_config is None or spec.is_oauth:
                 continue
             providers.append(
                 {
                     "name": spec.name,
                     "label": spec.label,
-                    "configured": bool(provider_config.api_key),
+                    "configured": _provider_configured_for_settings(spec, provider_config),
+                    "api_key_required": _provider_requires_api_key(spec),
                     "api_key_hint": _mask_secret_hint(provider_config.api_key),
                     "api_base": provider_config.api_base,
                     "default_api_base": spec.default_api_base or None,
@@ -862,7 +882,12 @@ class WebSocketChannel(BaseChannel):
             if find_by_name(provider) is None:
                 return _http_error(400, "unknown provider")
             provider_config = getattr(config.providers, provider, None)
-            if provider_config is None or not provider_config.api_key:
+            spec = find_by_name(provider)
+            if (
+                provider_config is None
+                or spec is None
+                or not _provider_configured_for_settings(spec, provider_config)
+            ):
                 return _http_error(400, "provider is not configured")
             if defaults.provider != provider:
                 defaults.provider = provider
@@ -885,7 +910,7 @@ class WebSocketChannel(BaseChannel):
         if not provider_name:
             return _http_error(400, "provider is required")
         spec = find_by_name(provider_name)
-        if spec is None or spec.is_oauth or spec.is_local:
+        if spec is None or spec.is_oauth:
             return _http_error(400, "unknown provider")
 
         config = load_config()
@@ -1581,6 +1606,7 @@ class WebSocketChannel(BaseChannel):
         if not conns:
             if (
                 msg.metadata.get("_progress")
+                or msg.metadata.get("_file_edit_events")
                 or msg.metadata.get("_turn_end")
                 or msg.metadata.get("_session_updated")
                 or msg.metadata.get("_goal_status")
@@ -1613,7 +1639,22 @@ class WebSocketChannel(BaseChannel):
             await self.send_turn_end(msg.chat_id, latency_ms=lat_i, goal_state=gs_blob)
             return
         if msg.metadata.get("_session_updated"):
-            await self.send_session_updated(msg.chat_id)
+            scope = msg.metadata.get("_session_update_scope")
+            await self.send_session_updated(
+                msg.chat_id,
+                scope=scope if isinstance(scope, str) else None,
+            )
+            return
+        if msg.metadata.get("_file_edit_events"):
+            payload: dict[str, Any] = {
+                "event": "file_edit",
+                "chat_id": msg.chat_id,
+                "edits": msg.metadata["_file_edit_events"],
+            }
+            self._try_append_webui_transcript(msg.chat_id, payload)
+            raw = json.dumps(payload, ensure_ascii=False)
+            for connection in conns:
+                await self._safe_send_to(connection, raw, label=" ")
             return
         text = msg.content
         payload: dict[str, Any] = {
@@ -1780,12 +1821,14 @@ class WebSocketChannel(BaseChannel):
         for connection in conns:
             await self._safe_send_to(connection, raw, label=" goal_status ")
 
-    async def send_session_updated(self, chat_id: str) -> None:
+    async def send_session_updated(self, chat_id: str, *, scope: str | None = None) -> None:
         """Notify clients that session metadata changed outside the main turn."""
         conns = list(self._subs.get(chat_id, ()))
         if not conns:
             return
         body: dict[str, Any] = {"event": "session_updated", "chat_id": chat_id}
+        if scope:
+            body["scope"] = scope
         raw = json.dumps(body, ensure_ascii=False)
         for connection in conns:
             await self._safe_send_to(connection, raw, label=" session_updated ")

nanobot/cli/commands.py

@@ -91,6 +91,8 @@ app = typer.Typer(
 
 console = Console()
 EXIT_COMMANDS = {"exit", "quit", "/exit", "/quit", ":q"}
+_REASONING_SENTENCE_ENDINGS = (".", "!", "?", "。", "！", "？")
+_REASONING_FLUSH_CHARS = 60
 
 # ---------------------------------------------------------------------------
 # CLI input: prompt_toolkit for editing, paste, history, and display
@@ -242,6 +244,35 @@ def _print_cli_progress_line(text: str, thinking: ThinkingSpinner | None, render
         target.print(f"  [dim]↳ {text}[/dim]")
 
 
+class _ReasoningBuffer:
+    def __init__(self) -> None:
+        self._text = ""
+
+    def add(self, text: str) -> str | None:
+        if not text:
+            return None
+        self._text += text
+        if self._should_flush(text):
+            return self.flush()
+        return None
+
+    def flush(self) -> str | None:
+        text = self._text.strip()
+        self._text = ""
+        return text or None
+
+    def clear(self) -> None:
+        self._text = ""
+
+    def _should_flush(self, text: str) -> bool:
+        stripped = text.rstrip()
+        return (
+            "\n" in text
+            or stripped.endswith(_REASONING_SENTENCE_ENDINGS)
+            or len(self._text) >= _REASONING_FLUSH_CHARS
+        )
+
+
 def _print_cli_reasoning(text: str, thinking: ThinkingSpinner | None, renderer: StreamRenderer | None = None) -> None:
     """Print reasoning/thinking content in a distinct style."""
     if not text.strip():
@@ -254,6 +285,16 @@ def _print_cli_reasoning(text: str, thinking: ThinkingSpinner | None, renderer:
         target.print(f"[dim italic]✻ {text}[/dim italic]")
 
 
+def _flush_cli_reasoning(
+    reasoning_buffer: _ReasoningBuffer,
+    thinking: ThinkingSpinner | None,
+    renderer: StreamRenderer | None = None,
+) -> None:
+    text = reasoning_buffer.flush()
+    if text:
+        _print_cli_reasoning(text, thinking, renderer)
+
+
 async def _print_interactive_progress_line(text: str, thinking: ThinkingSpinner | None, renderer: StreamRenderer | None = None) -> None:
     """Print an interactive progress line, pausing the spinner if needed."""
     if not text.strip():
@@ -272,6 +313,7 @@ async def _maybe_print_interactive_progress(
     thinking: ThinkingSpinner | None,
     channels_config: Any,
     renderer: StreamRenderer | None = None,
+    reasoning_buffer: _ReasoningBuffer | None = None,
 ) -> bool:
     metadata = msg.metadata or {}
     if metadata.get("_retry_wait"):
@@ -281,12 +323,24 @@ async def _maybe_print_interactive_progress(
     if not metadata.get("_progress"):
         return False
 
+    reasoning_buffer = reasoning_buffer or _ReasoningBuffer()
+
+    if metadata.get("_reasoning_end"):
+        if channels_config and not channels_config.show_reasoning:
+            reasoning_buffer.clear()
+        else:
+            _flush_cli_reasoning(reasoning_buffer, thinking, renderer)
+        return True
+
     is_tool_hint = metadata.get("_tool_hint", False)
     is_reasoning = metadata.get("_reasoning", False) or metadata.get("_reasoning_delta", False)
     if is_reasoning:
         if channels_config and not channels_config.show_reasoning:
+            reasoning_buffer.clear()
             return True
-        _print_cli_reasoning(msg.content, thinking, renderer)
+        text = reasoning_buffer.add(msg.content)
+        if text:
+            _print_cli_reasoning(text, thinking, renderer)
         return True
     if channels_config and is_tool_hint and not channels_config.send_tool_hints:
         return True
@@ -918,8 +972,7 @@ def _run_gateway(
     hb_cfg = config.gateway.heartbeat
     heartbeat = HeartbeatService(
         workspace=config.workspace_path,
-        provider=agent.provider,
-        model=agent.model,
+        llm_runtime=agent.llm_runtime,
         on_execute=on_heartbeat_execute,
         on_notify=on_heartbeat_notify,
         interval_s=hb_cfg.interval_s,
@@ -1111,12 +1164,25 @@ def agent(
     _thinking: ThinkingSpinner | None = None
 
     def _make_progress(renderer: StreamRenderer | None = None):
+        reasoning_buffer = _ReasoningBuffer()
+
         async def _cli_progress(content: str, *, tool_hint: bool = False, reasoning: bool = False, **_kwargs: Any) -> None:
             ch = agent_loop.channels_config
+
+            if _kwargs.get("reasoning_end"):
+                if ch and not ch.show_reasoning:
+                    reasoning_buffer.clear()
+                else:
+                    _flush_cli_reasoning(reasoning_buffer, _thinking, renderer)
+                return
+
             if reasoning:
                 if ch and not ch.show_reasoning:
+                    reasoning_buffer.clear()
                     return
-                _print_cli_reasoning(content, _thinking, renderer)
+                text = reasoning_buffer.add(content)
+                if text:
+                    _print_cli_reasoning(text, _thinking, renderer)
                 return
             if ch and tool_hint and not ch.send_tool_hints:
                 return
@@ -1187,6 +1253,7 @@ def agent(
             turn_done.set()
             turn_response: list[tuple[str, dict]] = []
             renderer: StreamRenderer | None = None
+            reasoning_buffer = _ReasoningBuffer()
 
             async def _consume_outbound():
                 while True:
@@ -1212,6 +1279,7 @@ def agent(
                             renderer,
                             agent_loop.channels_config,
                             renderer,
+                            reasoning_buffer,
                         ):
                             continue
 
@@ -1252,6 +1320,7 @@ def agent(
 
                         turn_done.clear()
                         turn_response.clear()
+                        reasoning_buffer.clear()
                         renderer = StreamRenderer(
                             render_markdown=markdown,
                             bot_name=config.agents.defaults.bot_name,

nanobot/cli/onboard.py

@@ -22,7 +22,7 @@ from nanobot.cli.models import (
     get_model_suggestions,
 )
 from nanobot.config.loader import get_config_path, load_config
-from nanobot.config.schema import Config
+from nanobot.config.schema import Config, ModelPresetConfig
 
 console = Console()
 
@@ -49,6 +49,10 @@ _SELECT_FIELD_HINTS: dict[str, tuple[list[str], str]] = {
 
 _BACK_PRESSED = object()  # Sentinel value for back navigation
 
+# Cache of model-preset names populated at runtime so that field handlers can
+# offer existing presets as choices (e.g. AgentDefaults.model_preset).
+_MODEL_PRESET_CACHE: set[str] = set()
+
 
 def _get_questionary():
     """Return questionary or raise a clear error when wizard deps are unavailable."""
@@ -588,9 +592,102 @@ def _handle_context_window_field(
         setattr(working_model, field_name, new_value)
 
 
+def _handle_model_preset_field(
+    working_model: BaseModel, field_name: str, field_display: str, current_value: Any
+) -> None:
+    """Handle the 'model_preset' field with a list of existing presets."""
+    preset_names = sorted(_MODEL_PRESET_CACHE)
+    choices = ["(clear/unset)"] + preset_names
+    default_choice = str(current_value) if current_value else "(clear/unset)"
+    new_value = _select_with_back(field_display, choices, default=default_choice)
+    if new_value is _BACK_PRESSED:
+        return
+    if new_value == "(clear/unset)":
+        setattr(working_model, field_name, None)
+    elif new_value is not None:
+        setattr(working_model, field_name, new_value)
+
+
+def _handle_provider_field(
+    working_model: BaseModel, field_name: str, field_display: str, current_value: Any
+) -> None:
+    """Handle the 'provider' field with a list of registered providers."""
+    provider_names = sorted(_get_provider_names().keys())
+    choices = ["auto"] + provider_names
+    default_choice = str(current_value) if current_value else "auto"
+    new_value = _select_with_back(field_display, choices, default=default_choice)
+    if new_value is _BACK_PRESSED:
+        return
+    if new_value is not None:
+        setattr(working_model, field_name, new_value)
+
+
+def _handle_fallback_models_field(
+    working_model: BaseModel, field_name: str, field_display: str, current_value: Any
+) -> None:
+    """Handle the 'fallback_models' field with preset-aware list management."""
+    from nanobot.config.schema import InlineFallbackConfig
+
+    items: list[Any] = list(current_value) if isinstance(current_value, list) else []
+    preset_names = sorted(_MODEL_PRESET_CACHE)
+
+    while True:
+        console.clear()
+        console.print(f"[bold]{field_display}[/bold]")
+        if items:
+            for idx, item in enumerate(items, 1):
+                if isinstance(item, InlineFallbackConfig):
+                    console.print(f"  {idx}. {item.model} ({item.provider}) [inline]")
+                else:
+                    console.print(f"  {idx}. {item}")
+        else:
+            console.print("  [dim](empty)[/dim]")
+        console.print()
+
+        choices = ["[+] Add preset"]
+        if items:
+            choices.append("[-] Remove last")
+            choices.append("[X] Clear all")
+        choices.append("[Done]")
+        choices.append("<- Back")
+
+        answer = _get_questionary().select(
+            "Manage fallback models:",
+            choices=choices,
+            qmark=">",
+        ).ask()
+
+        if answer is None or answer == "<- Back":
+            return
+        if answer == "[Done]":
+            setattr(working_model, field_name, items)
+            return
+        if answer == "[+] Add preset":
+            if not preset_names:
+                console.print("[yellow]! No presets defined yet.[/yellow]")
+                _get_questionary().press_any_key_to_continue().ask()
+                continue
+            add_choices = [p for p in preset_names if p not in items]
+            if not add_choices:
+                console.print("[yellow]! All presets already added.[/yellow]")
+                _get_questionary().press_any_key_to_continue().ask()
+                continue
+            picked = _select_with_back("Select preset:", add_choices)
+            if picked is _BACK_PRESSED or picked is None:
+                continue
+            items.append(picked)
+        elif answer == "[-] Remove last" and items:
+            items.pop()
+        elif answer == "[X] Clear all" and items:
+            items.clear()
+
+
 _FIELD_HANDLERS: dict[str, Any] = {
     "model": _handle_model_field,
     "context_window_tokens": _handle_context_window_field,
+    "model_preset": _handle_model_preset_field,
+    "provider": _handle_provider_field,
+    "fallback_models": _handle_fallback_models_field,
 }
 
 
@@ -757,6 +854,116 @@ def _try_auto_fill_context_window(model: BaseModel, new_model_name: str) -> None
         console.print("[dim](i) Could not auto-fill context window (model not in database)[/dim]")
 
 
+# --- Model Preset Configuration ---
+
+
+def _sync_preset_cache(config: Config) -> None:
+    """Synchronise the module-level preset name cache from config."""
+    _MODEL_PRESET_CACHE.clear()
+    _MODEL_PRESET_CACHE.update(config.model_presets.keys())
+
+
+def _configure_model_presets(config: Config) -> None:
+    """Configure model presets (CRUD)."""
+    _sync_preset_cache(config)
+
+    def get_preset_choices() -> list[str]:
+        choices: list[str] = []
+        for name, preset in config.model_presets.items():
+            choices.append(f"{name} ({preset.model})")
+        choices.append("[+] Add new preset")
+        choices.append("<- Back")
+        return choices
+
+    last_preset_name: str | None = None
+    while True:
+        try:
+            console.clear()
+            _show_section_header(
+                "Model Presets",
+                "Create, edit or delete named model presets for quick switching",
+            )
+            choices = get_preset_choices()
+            default_choice = None
+            if last_preset_name:
+                for c in choices:
+                    if c.startswith(last_preset_name + " ("):
+                        default_choice = c
+                        break
+            answer = _select_with_back(
+                "Select preset:", choices, default=default_choice
+            )
+
+            if answer is _BACK_PRESSED or answer is None or answer == "<- Back":
+                break
+
+            assert isinstance(answer, str)
+
+            if answer == "[+] Add new preset":
+                name_input = _get_questionary().text(
+                    "Preset name:",
+                    validate=lambda t: True if t and t.strip() else "Name cannot be empty",
+                ).ask()
+                if not name_input:
+                    continue
+                name = name_input.strip()
+                if name in config.model_presets:
+                    console.print(f"[yellow]! Preset '{name}' already exists[/yellow]")
+                    _pause()
+                    continue
+                if name == "default":
+                    console.print("[yellow]! 'default' is reserved (auto-generated from Agent Settings)[/yellow]")
+                    _pause()
+                    continue
+                new_preset = ModelPresetConfig(model="")
+                updated = _configure_pydantic_model(new_preset, f"New Preset: {name}")
+                if updated is not None:
+                    config.model_presets[name] = updated
+                    _sync_preset_cache(config)
+                    last_preset_name = name
+                continue
+
+            # Editing / deleting an existing preset
+            preset_name = answer.split(" (", 1)[0]
+            preset = config.model_presets.get(preset_name)
+            if preset is None:
+                continue
+
+            last_preset_name = preset_name
+
+            choices = ["Edit", "Cancel"]
+            if preset_name != "default":
+                choices.insert(1, "Delete")
+            action = _select_with_back(
+                f"Preset: {preset_name}",
+                choices,
+                default="Edit",
+            )
+            if action is _BACK_PRESSED or action == "Cancel" or action is None:
+                continue
+
+            if action == "Delete":
+                confirm = _get_questionary().confirm(
+                    f"Delete preset '{preset_name}'?",
+                    default=False,
+                ).ask()
+                if confirm:
+                    del config.model_presets[preset_name]
+                    _sync_preset_cache(config)
+                    last_preset_name = None
+                continue
+
+            if action == "Edit":
+                updated = _configure_pydantic_model(preset, f"Edit Preset: {preset_name}")
+                if updated is not None:
+                    config.model_presets[preset_name] = updated
+                    _sync_preset_cache(config)
+
+        except KeyboardInterrupt:
+            console.print("\n[dim]Returning to main menu...[/dim]")
+            break
+
+
 # --- Provider Configuration ---
 
 
@@ -1043,6 +1250,12 @@ def _show_summary(config: Config) -> None:
         channel_rows.append((display, status))
     _print_summary_panel(channel_rows, "Chat Channels")
 
+    # Model Presets
+    preset_rows = []
+    for name, preset in config.model_presets.items():
+        preset_rows.append((name, f"{preset.model} (ctx={preset.context_window_tokens})"))
+    _print_summary_panel(preset_rows, "Model Presets")
+
     # Settings sections
     for title, model in [
         ("Agent Settings", config.agents.defaults),
@@ -1112,6 +1325,7 @@ def run_onboard(initial_config: Config | None = None) -> OnboardResult:
 
     original_config = base_config.model_copy(deep=True)
     config = base_config.model_copy(deep=True)
+    _sync_preset_cache(config)
 
     last_main_choice: str | None = None
     while True:
@@ -1123,6 +1337,7 @@ def run_onboard(initial_config: Config | None = None) -> OnboardResult:
                 "What would you like to configure?",
                 choices=[
                     "[P] LLM Provider",
+                    "[M] Model Presets",
                     "[C] Chat Channel",
                     "[H] Channel Common",
                     "[A] Agent Settings",
@@ -1149,6 +1364,7 @@ def run_onboard(initial_config: Config | None = None) -> OnboardResult:
 
         _menu_dispatch = {
             "[P] LLM Provider": lambda: _configure_providers(config),
+            "[M] Model Presets": lambda: _configure_model_presets(config),
             "[C] Chat Channel": lambda: _configure_channels(config),
             "[H] Channel Common": lambda: _configure_general_settings(config, "Channel Common"),
             "[A] Agent Settings": lambda: _configure_general_settings(config, "Agent Settings"),

nanobot/heartbeat/service.py

@@ -4,12 +4,12 @@ from __future__ import annotations
 
 import asyncio
 from pathlib import Path
-from typing import TYPE_CHECKING, Any, Callable, Coroutine
+from typing import Any, Callable, Coroutine
 
 from loguru import logger
 
-if TYPE_CHECKING:
-    from nanobot.providers.base import LLMProvider
+from nanobot.providers.base import LLMProvider
+from nanobot.utils.llm_runtime import LLMRuntimeResolver, static_llm_runtime
 
 _HEARTBEAT_TOOL = [
     {
@@ -53,17 +53,21 @@ class HeartbeatService:
     def __init__(
         self,
         workspace: Path,
-        provider: LLMProvider,
-        model: str,
+        provider: LLMProvider | None = None,
+        model: str | None = None,
         on_execute: Callable[[str], Coroutine[Any, Any, str]] | None = None,
         on_notify: Callable[[str], Coroutine[Any, Any, None]] | None = None,
         interval_s: int = 30 * 60,
         enabled: bool = True,
         timezone: str | None = None,
+        llm_runtime: LLMRuntimeResolver | None = None,
     ):
         self.workspace = workspace
-        self.provider = provider
-        self.model = model
+        if llm_runtime is None:
+            if provider is None or model is None:
+                raise ValueError("HeartbeatService requires either llm_runtime or provider/model")
+            llm_runtime = static_llm_runtime(provider, model)
+        self._llm_runtime = llm_runtime
         self.on_execute = on_execute
         self.on_notify = on_notify
         self.interval_s = interval_s
@@ -91,7 +95,9 @@ class HeartbeatService:
         """
         from nanobot.utils.helpers import current_time_str
 
-        response = await self.provider.chat_with_retry(
+        llm = self._llm_runtime()
+
+        response = await llm.provider.chat_with_retry(
             messages=[
                 {"role": "system", "content": "You are a heartbeat agent. Call the heartbeat tool to report your decision."},
                 {"role": "user", "content": (
@@ -101,7 +107,7 @@ class HeartbeatService:
                 )},
             ],
             tools=_HEARTBEAT_TOOL,
-            model=self.model,
+            model=llm.model,
         )
 
         if not response.should_execute_tools:
@@ -214,8 +220,9 @@ class HeartbeatService:
                     )
                     return
 
+                llm = self._llm_runtime()
                 should_notify = await evaluate_response(
-                    response, tasks, self.provider, self.model,
+                    response, tasks, llm.provider, llm.model,
                 )
                 if should_notify and self.on_notify:
                     logger.info("Heartbeat: completed, delivering response")

nanobot/providers/registry.py

@@ -396,7 +396,7 @@ PROVIDERS: tuple[ProviderSpec, ...] = (
         name="vllm",
         keywords=("vllm",),
         env_key="HOSTED_VLLM_API_KEY",
-        display_name="vLLM/Local",
+        display_name="vLLM",
         backend="openai_compat",
         is_local=True,
     ),

nanobot/utils/file_edit_events.py

@@ -0,0 +1,311 @@
+"""File-edit activity helpers for WebUI progress events."""
+
+from __future__ import annotations
+
+import difflib
+import json
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Any
+
+
+TRACKED_FILE_EDIT_TOOLS = frozenset({"write_file", "edit_file", "notebook_edit"})
+_MAX_SNAPSHOT_BYTES = 2 * 1024 * 1024
+
+
+@dataclass(slots=True)
+class FileSnapshot:
+    path: Path
+    exists: bool
+    text: str | None
+    unreadable: bool = False
+    binary: bool = False
+    oversized: bool = False
+
+    @property
+    def countable(self) -> bool:
+        return (
+            self.text is not None
+            and not self.binary
+            and not self.oversized
+            and not self.unreadable
+        )
+
+
+@dataclass(slots=True)
+class FileEditTracker:
+    call_id: str
+    tool: str
+    path: Path
+    display_path: str
+    before: FileSnapshot
+
+
+def is_file_edit_tool(tool_name: str | None) -> bool:
+    return bool(tool_name) and tool_name in TRACKED_FILE_EDIT_TOOLS
+
+
+def resolve_file_edit_path(
+    tool: Any,
+    workspace: Path | None,
+    params: dict[str, Any] | None,
+) -> Path | None:
+    """Resolve the target file path after tool argument preparation."""
+    if not isinstance(params, dict):
+        return None
+    raw_path = params.get("path")
+    if not isinstance(raw_path, str) or not raw_path.strip():
+        return None
+    resolver = getattr(tool, "_resolve", None)
+    if callable(resolver):
+        try:
+            resolved = resolver(raw_path)
+            if isinstance(resolved, Path):
+                return resolved
+            if resolved:
+                return Path(resolved)
+        except Exception:
+            return None
+    if workspace is None:
+        return Path(raw_path).expanduser().resolve()
+    return (workspace / raw_path).expanduser().resolve()
+
+
+def display_file_edit_path(path: Path, workspace: Path | None) -> str:
+    if workspace is not None:
+        try:
+            return path.resolve().relative_to(workspace.resolve()).as_posix()
+        except Exception:
+            pass
+    return path.as_posix()
+
+
+def read_file_snapshot(path: Path, *, max_bytes: int = _MAX_SNAPSHOT_BYTES) -> FileSnapshot:
+    try:
+        if not path.exists() or not path.is_file():
+            return FileSnapshot(path=path, exists=False, text="")
+        size = path.stat().st_size
+        if size > max_bytes:
+            return FileSnapshot(path=path, exists=True, text=None, oversized=True)
+        raw = path.read_bytes()
+    except OSError:
+        return FileSnapshot(path=path, exists=path.exists(), text=None, unreadable=True)
+    if b"\x00" in raw:
+        return FileSnapshot(path=path, exists=True, text=None, binary=True)
+    try:
+        text = raw.decode("utf-8")
+    except UnicodeDecodeError:
+        return FileSnapshot(path=path, exists=True, text=None, binary=True)
+    return FileSnapshot(path=path, exists=True, text=text.replace("\r\n", "\n"))
+
+
+def line_diff_stats(before: str | None, after: str | None) -> tuple[int, int]:
+    """Return ``(added, deleted)`` for a UTF-8 text line-level diff."""
+    if before is None or after is None:
+        return 0, 0
+    before_lines = before.replace("\r\n", "\n").splitlines()
+    after_lines = after.replace("\r\n", "\n").splitlines()
+    added = 0
+    deleted = 0
+    matcher = difflib.SequenceMatcher(a=before_lines, b=after_lines, autojunk=False)
+    for tag, i1, i2, j1, j2 in matcher.get_opcodes():
+        if tag == "equal":
+            continue
+        if tag in ("replace", "delete"):
+            deleted += i2 - i1
+        if tag in ("replace", "insert"):
+            added += j2 - j1
+    return added, deleted
+
+
+def prepare_file_edit_tracker(
+    *,
+    call_id: str,
+    tool_name: str,
+    tool: Any,
+    workspace: Path | None,
+    params: dict[str, Any] | None,
+) -> FileEditTracker | None:
+    if not is_file_edit_tool(tool_name):
+        return None
+    path = resolve_file_edit_path(tool, workspace, params)
+    if path is None:
+        return None
+    before = read_file_snapshot(path)
+    return FileEditTracker(
+        call_id=str(call_id or ""),
+        tool=tool_name,
+        path=path,
+        display_path=display_file_edit_path(path, workspace),
+        before=before,
+    )
+
+
+def build_file_edit_start_event(
+    tracker: FileEditTracker,
+    params: dict[str, Any] | None,
+) -> dict[str, Any]:
+    predicted_after = _predict_after_text(tracker.tool, params or {}, tracker.before)
+    if tracker.before.countable and predicted_after is not None:
+        added, deleted = line_diff_stats(tracker.before.text, predicted_after)
+    else:
+        added, deleted = 0, 0
+    return _event_payload(
+        tracker,
+        phase="start",
+        status="editing",
+        added=added,
+        deleted=deleted,
+        approximate=True,
+    )
+
+
+def build_file_edit_end_event(tracker: FileEditTracker) -> dict[str, Any]:
+    after = read_file_snapshot(tracker.path)
+    if tracker.before.countable and after.countable:
+        added, deleted = line_diff_stats(tracker.before.text, after.text)
+    else:
+        added, deleted = 0, 0
+    return _event_payload(
+        tracker,
+        phase="end",
+        status="done",
+        added=added,
+        deleted=deleted,
+        approximate=False,
+        binary=after.binary or after.oversized or after.unreadable,
+    )
+
+
+def build_file_edit_error_event(tracker: FileEditTracker, error: str | None = None) -> dict[str, Any]:
+    payload = _event_payload(
+        tracker,
+        phase="error",
+        status="error",
+        added=0,
+        deleted=0,
+        approximate=False,
+    )
+    if error:
+        payload["error"] = error.strip()[:240]
+    return payload
+
+
+def _event_payload(
+    tracker: FileEditTracker,
+    *,
+    phase: str,
+    status: str,
+    added: int,
+    deleted: int,
+    approximate: bool,
+    binary: bool = False,
+) -> dict[str, Any]:
+    payload: dict[str, Any] = {
+        "version": 1,
+        "call_id": tracker.call_id,
+        "tool": tracker.tool,
+        "path": tracker.display_path,
+        "phase": phase,
+        "added": max(0, int(added)),
+        "deleted": max(0, int(deleted)),
+        "approximate": bool(approximate),
+        "status": status,
+    }
+    if binary:
+        payload["binary"] = True
+    return payload
+
+
+def _predict_after_text(
+    tool_name: str,
+    params: dict[str, Any],
+    before: FileSnapshot,
+) -> str | None:
+    if not before.countable:
+        return None
+    before_text = before.text or ""
+    if tool_name == "write_file":
+        content = params.get("content")
+        return content if isinstance(content, str) else ""
+    if tool_name == "edit_file":
+        old_text = params.get("old_text")
+        new_text = params.get("new_text")
+        if not isinstance(old_text, str) or not isinstance(new_text, str):
+            return None
+        replace_all = bool(params.get("replace_all"))
+        if old_text == "":
+            return new_text if not before.exists else before_text
+        if old_text in before_text:
+            if replace_all:
+                return before_text.replace(old_text, new_text)
+            return before_text.replace(old_text, new_text, 1)
+        return None
+    if tool_name == "notebook_edit":
+        return _predict_notebook_after_text(params, before_text)
+    return None
+
+
+def _predict_notebook_after_text(params: dict[str, Any], before_text: str) -> str | None:
+    try:
+        nb = json.loads(before_text) if before_text.strip() else _empty_notebook()
+    except Exception:
+        return None
+    cells = nb.get("cells")
+    if not isinstance(cells, list):
+        return None
+    try:
+        cell_index = int(params.get("cell_index", 0))
+    except (TypeError, ValueError):
+        return None
+    new_source = params.get("new_source")
+    source = new_source if isinstance(new_source, str) else ""
+    cell_type = params.get("cell_type") if params.get("cell_type") in ("code", "markdown") else "code"
+    mode = params.get("edit_mode") if params.get("edit_mode") in ("replace", "insert", "delete") else "replace"
+    if mode == "delete":
+        if 0 <= cell_index < len(cells):
+            cells.pop(cell_index)
+        else:
+            return None
+    elif mode == "insert":
+        insert_at = min(max(cell_index + 1, 0), len(cells))
+        cells.insert(insert_at, _new_notebook_cell(source, str(cell_type)))
+    else:
+        if not (0 <= cell_index < len(cells)):
+            return None
+        cell = cells[cell_index]
+        if not isinstance(cell, dict):
+            return None
+        cell["source"] = source
+        cell["cell_type"] = cell_type
+        if cell_type == "code":
+            cell.setdefault("outputs", [])
+            cell.setdefault("execution_count", None)
+        else:
+            cell.pop("outputs", None)
+            cell.pop("execution_count", None)
+    nb["cells"] = cells
+    try:
+        return json.dumps(nb, indent=1, ensure_ascii=False)
+    except Exception:
+        return None
+
+
+def _empty_notebook() -> dict[str, Any]:
+    return {
+        "nbformat": 4,
+        "nbformat_minor": 5,
+        "metadata": {
+            "kernelspec": {"display_name": "Python 3", "language": "python", "name": "python3"},
+            "language_info": {"name": "python"},
+        },
+        "cells": [],
+    }
+
+
+def _new_notebook_cell(source: str, cell_type: str) -> dict[str, Any]:
+    cell: dict[str, Any] = {"cell_type": cell_type, "source": source, "metadata": {}}
+    if cell_type == "code":
+        cell["outputs"] = []
+        cell["execution_count"] = None
+    return cell

nanobot/utils/llm_runtime.py

@@ -0,0 +1,22 @@
+"""Small helpers for passing the active LLM provider/model together."""
+
+from __future__ import annotations
+
+from collections.abc import Callable
+from dataclasses import dataclass
+
+from nanobot.providers.base import LLMProvider
+
+
+@dataclass(frozen=True)
+class LLMRuntime:
+    provider: LLMProvider
+    model: str
+
+
+LLMRuntimeResolver = Callable[[], LLMRuntime]
+
+
+def static_llm_runtime(provider: LLMProvider, model: str) -> LLMRuntimeResolver:
+    runtime = LLMRuntime(provider=provider, model=model)
+    return lambda: runtime

nanobot/utils/progress_events.py

@@ -10,13 +10,21 @@ from nanobot.agent.hook import AgentHookContext
 
 
 def on_progress_accepts_tool_events(cb: Callable[..., Any]) -> bool:
+    return _on_progress_accepts(cb, "tool_events")
+
+
+def on_progress_accepts_file_edit_events(cb: Callable[..., Any]) -> bool:
+    return _on_progress_accepts(cb, "file_edit_events")
+
+
+def _on_progress_accepts(cb: Callable[..., Any], name: str) -> bool:
     try:
         sig = inspect.signature(cb)
     except (TypeError, ValueError):
         return False
     if any(p.kind == inspect.Parameter.VAR_KEYWORD for p in sig.parameters.values()):
         return True
-    return "tool_events" in sig.parameters
+    return name in sig.parameters
 
 
 async def invoke_on_progress(
@@ -32,6 +40,15 @@ async def invoke_on_progress(
     await on_progress(content, tool_hint=tool_hint)
 
 
+async def invoke_file_edit_progress(
+    on_progress: Callable[..., Awaitable[None]],
+    file_edit_events: list[dict[str, Any]],
+) -> None:
+    if not file_edit_events or not on_progress_accepts_file_edit_events(on_progress):
+        return
+    await on_progress("", file_edit_events=file_edit_events)
+
+
 def build_tool_event_start_payload(tool_call: Any) -> dict[str, Any]:
     return {
         "version": 1,

nanobot/utils/webui_titles.py

@@ -1,138 +0,0 @@
-"""Helpers for WebUI chat title generation."""
-
-from __future__ import annotations
-
-import re
-from typing import Any
-
-from loguru import logger
-
-from nanobot.providers.base import LLMProvider
-from nanobot.session.manager import Session, SessionManager
-from nanobot.utils.helpers import truncate_text
-
-WEBUI_SESSION_METADATA_KEY = "webui"
-WEBUI_TITLE_METADATA_KEY = "title"
-WEBUI_TITLE_USER_EDITED_METADATA_KEY = "title_user_edited"
-TITLE_MAX_CHARS = 60
-
-
-def mark_webui_session(session: Session, metadata: dict[str, Any]) -> bool:
-    """Persist a WebUI marker only when the inbound websocket frame opted in."""
-    if metadata.get(WEBUI_SESSION_METADATA_KEY) is not True:
-        return False
-    session.metadata[WEBUI_SESSION_METADATA_KEY] = True
-    return True
-
-
-def clean_generated_title(raw: str | None) -> str:
-    text = (raw or "").strip()
-    if not text:
-        return ""
-    text = re.sub(r"^\s*(title|标题)\s*[:：]\s*", "", text, flags=re.IGNORECASE)
-    text = text.strip().strip("\"'`“”‘’")
-    text = re.sub(r"\s+", " ", text).strip()
-    text = text.rstrip("。.!！?？,，;；:")
-    if len(text) > TITLE_MAX_CHARS:
-        text = text[: TITLE_MAX_CHARS - 1].rstrip() + "…"
-    return text
-
-
-def _title_inputs(session: Session) -> tuple[str, str]:
-    user_text = ""
-    assistant_text = ""
-    for message in session.messages:
-        role = message.get("role")
-        content = message.get("content")
-        if not isinstance(content, str) or not content.strip():
-            continue
-        if role == "user" and not user_text:
-            user_text = content.strip()
-        elif role == "assistant" and not assistant_text:
-            assistant_text = content.strip()
-        if user_text and assistant_text:
-            break
-    return user_text, assistant_text
-
-
-async def maybe_generate_webui_title(
-    *,
-    sessions: SessionManager,
-    session_key: str,
-    provider: LLMProvider,
-    model: str,
-) -> bool:
-    """Generate and persist a short title for WebUI-owned sessions only."""
-    session = sessions.get_or_create(session_key)
-    if session.metadata.get(WEBUI_SESSION_METADATA_KEY) is not True:
-        return False
-    if session.metadata.get(WEBUI_TITLE_USER_EDITED_METADATA_KEY) is True:
-        return False
-    current_title = session.metadata.get(WEBUI_TITLE_METADATA_KEY)
-    if isinstance(current_title, str) and current_title.strip():
-        return False
-
-    user_text, assistant_text = _title_inputs(session)
-    if not user_text:
-        return False
-
-    prompt = (
-        "Generate a concise title for this chat.\n"
-        "Rules:\n"
-        "- Use the same language as the user when practical.\n"
-        "- 3 to 8 words.\n"
-        "- No quotes.\n"
-        "- No punctuation at the end.\n"
-        "- Return only the title.\n\n"
-        f"User: {truncate_text(user_text, 1_000)}"
-    )
-    if assistant_text:
-        prompt += f"\nAssistant: {truncate_text(assistant_text, 1_000)}"
-
-    try:
-        response = await provider.chat_with_retry(
-            [
-                {
-                    "role": "system",
-                    "content": (
-                        "You write short, neutral chat titles. "
-                        "Return only the title text."
-                    ),
-                },
-                {"role": "user", "content": prompt},
-            ],
-            tools=None,
-            model=model,
-            max_tokens=32,
-            temperature=0.2,
-            retry_mode="standard",
-        )
-    except Exception:
-        logger.debug("Failed to generate webui session title for {}", session_key, exc_info=True)
-        return False
-
-    title = clean_generated_title(response.content)
-    if not title or title.lower().startswith("error"):
-        return False
-    session.metadata[WEBUI_TITLE_METADATA_KEY] = title
-    sessions.save(session)
-    return True
-
-
-async def maybe_generate_webui_title_after_turn(
-    *,
-    channel: str,
-    metadata: dict[str, Any],
-    sessions: SessionManager,
-    session_key: str,
-    provider: LLMProvider,
-    model: str,
-) -> bool:
-    if channel != "websocket" or metadata.get(WEBUI_SESSION_METADATA_KEY) is not True:
-        return False
-    return await maybe_generate_webui_title(
-        sessions=sessions,
-        session_key=session_key,
-        provider=provider,
-        model=model,
-    )

nanobot/utils/webui_transcript.py

@@ -125,11 +125,25 @@ def replay_transcript_to_ui_messages(
     buffer_message_id: str | None = None
     buffer_parts: list[str] = []
     suppress_until_turn_end = False
+    active_activity_segment_id: str | None = None
+    active_file_edit_segment_id: str | None = None
+    activity_segment_counter = 0
     _ts_base = int(time.time() * 1000)
 
     def _new_id(prefix: str, idx: int) -> str:
         return f"{prefix}-{idx}-{uuid.uuid4().hex[:8]}"
 
+    def _new_activity_segment(*, activate: bool = True) -> str:
+        nonlocal active_activity_segment_id, activity_segment_counter
+        activity_segment_counter += 1
+        segment_id = f"activity-{activity_segment_counter}"
+        if activate:
+            active_activity_segment_id = segment_id
+        return segment_id
+
+    def _ensure_activity_segment() -> str:
+        return active_activity_segment_id or _new_activity_segment()
+
     def attach_reasoning_chunk(prev: list[dict[str, Any]], chunk: str, idx: int) -> None:
         for i in range(len(prev) - 1, -1, -1):
             candidate = prev[i]
@@ -151,12 +165,19 @@ def replay_transcript_to_ui_messages(
                     **candidate,
                     "reasoning": (str(candidate.get("reasoning") or "")) + chunk,
                     "reasoningStreaming": True,
+                    "activitySegmentId": candidate.get("activitySegmentId") or _ensure_activity_segment(),
                 }
                 return
             if not has_answer and candidate.get("isStreaming"):
-                prev[i] = {**candidate, "reasoning": chunk, "reasoningStreaming": True}
+                prev[i] = {
+                    **candidate,
+                    "reasoning": chunk,
+                    "reasoningStreaming": True,
+                    "activitySegmentId": candidate.get("activitySegmentId") or _ensure_activity_segment(),
+                }
                 return
             break
+        segment = _ensure_activity_segment()
         prev.append(
             {
                 "id": _new_id("as", idx),
@@ -165,6 +186,7 @@ def replay_transcript_to_ui_messages(
                 "isStreaming": True,
                 "reasoning": chunk,
                 "reasoningStreaming": True,
+                "activitySegmentId": segment,
                 "createdAt": _ts_base + idx,
             },
         )
@@ -221,6 +243,7 @@ def replay_transcript_to_ui_messages(
                 return
 
     def absorb_complete(extra: dict[str, Any], idx: int) -> None:
+        nonlocal active_activity_segment_id
         last = messages[-1] if messages else None
         if last and is_reasoning_only_placeholder(last):
             messages[-1] = {
@@ -238,10 +261,76 @@ def replay_transcript_to_ui_messages(
                     **extra,
                 },
             )
+        active_activity_segment_id = None
+
+    def _file_edit_key(edit: dict[str, Any]) -> str:
+        return "|".join(
+            str(edit.get(k) or "")
+            for k in ("call_id", "tool", "path")
+        )
+
+    def upsert_file_edits(edits: list[dict[str, Any]], idx: int) -> None:
+        nonlocal active_file_edit_segment_id
+        if not edits:
+            return
+        last = messages[-1] if messages else None
+        if (
+            active_file_edit_segment_id
+            and last
+            and last.get("kind") == "trace"
+            and last.get("fileEdits")
+        ):
+            segment = active_file_edit_segment_id
+        else:
+            segment = _new_activity_segment(activate=False)
+            active_file_edit_segment_id = segment
+        if not (
+            last
+            and last.get("kind") == "trace"
+            and not last.get("isStreaming")
+            and last.get("fileEdits")
+            and last.get("activitySegmentId") == segment
+        ):
+            messages.append(
+                {
+                    "id": _new_id("tr", idx),
+                    "role": "tool",
+                    "kind": "trace",
+                    "content": "",
+                    "traces": [],
+                    "fileEdits": [],
+                    "activitySegmentId": segment,
+                    "createdAt": _ts_base + idx,
+                },
+            )
+            last = messages[-1]
+        existing = list(last.get("fileEdits") or [])
+        index_by_key = {
+            _file_edit_key(edit): pos
+            for pos, edit in enumerate(existing)
+            if isinstance(edit, dict)
+        }
+        for edit in edits:
+            if not isinstance(edit, dict):
+                continue
+            key = _file_edit_key(edit)
+            if key in index_by_key:
+                pos = index_by_key[key]
+                existing[pos] = {**existing[pos], **edit}
+            else:
+                index_by_key[key] = len(existing)
+                existing.append(dict(edit))
+        messages[-1] = {
+            **last,
+            "fileEdits": existing,
+            "activitySegmentId": last.get("activitySegmentId") or segment,
+        }
 
     for idx, rec in enumerate(lines):
         ev = rec.get("event")
         if ev == "user":
+            active_activity_segment_id = None
+            active_file_edit_segment_id = None
             text = rec.get("text")
             text_s = text if isinstance(text, str) else ""
             media_paths = rec.get("media_paths")
@@ -264,6 +353,12 @@ def replay_transcript_to_ui_messages(
             messages.append(row)
             continue
 
+        if ev == "file_edit":
+            raw_edits = rec.get("edits")
+            if isinstance(raw_edits, list):
+                upsert_file_edits([e for e in raw_edits if isinstance(e, dict)], idx)
+            continue
+
         if ev == "delta":
             if suppress_until_turn_end:
                 continue
@@ -338,14 +433,21 @@ def replay_transcript_to_ui_messages(
                 trace_lines = structured if structured else ([text] if isinstance(text, str) and text else [])
                 if not trace_lines:
                     continue
+                segment = _ensure_activity_segment()
                 last = messages[-1] if messages else None
-                if last and last.get("kind") == "trace" and not last.get("isStreaming"):
+                if (
+                    last
+                    and last.get("kind") == "trace"
+                    and not last.get("isStreaming")
+                    and (last.get("activitySegmentId") in (None, segment))
+                ):
                     prev_traces = list(last.get("traces") or [last.get("content")])
                     merged_traces = prev_traces + trace_lines
                     messages[-1] = {
                         **last,
                         "traces": merged_traces,
                         "content": trace_lines[-1],
+                        "activitySegmentId": last.get("activitySegmentId") or segment,
                     }
                 else:
                     messages.append(
@@ -355,6 +457,7 @@ def replay_transcript_to_ui_messages(
                             "kind": "trace",
                             "content": trace_lines[-1],
                             "traces": trace_lines,
+                            "activitySegmentId": segment,
                             "createdAt": _ts_base + idx,
                         },
                     )
@@ -389,6 +492,8 @@ def replay_transcript_to_ui_messages(
 
         if ev == "turn_end":
             suppress_until_turn_end = False
+            active_activity_segment_id = None
+            active_file_edit_segment_id = None
             for i, m in enumerate(messages):
                 if m.get("isStreaming"):
                     messages[i] = {**m, "isStreaming": False}

nanobot/utils/webui_turn_helpers.py

@@ -6,17 +6,163 @@ AgentLoop uses these without importing a concrete channel plugin; only
 
 from __future__ import annotations
 
+import re
 import time
+from collections.abc import Awaitable, Callable
+from dataclasses import dataclass, field
 from typing import Any
 
+from loguru import logger
+
 from nanobot.bus.events import InboundMessage, OutboundMessage
 from nanobot.bus.queue import MessageBus
+from nanobot.providers.base import LLMProvider
+from nanobot.session.goal_state import goal_state_ws_blob
+from nanobot.session.manager import Session, SessionManager
+from nanobot.utils.helpers import truncate_text
+from nanobot.utils.llm_runtime import LLMRuntime
+
+WEBUI_SESSION_METADATA_KEY = "webui"
+WEBUI_TITLE_METADATA_KEY = "title"
+WEBUI_TITLE_USER_EDITED_METADATA_KEY = "title_user_edited"
+TITLE_MAX_CHARS = 60
+TITLE_GENERATION_MAX_TOKENS = 96
+TITLE_GENERATION_REASONING_EFFORT = "none"
 
 # Wall-clock turn start per ``chat_id`` (websocket only). Survives browser refresh while the
 # gateway process stays up; cleared on idle/stop and implicitly dropped on restart.
 _WEBSOCKET_TURN_WALL_STARTED_AT: dict[str, float] = {}
 
 
+def mark_webui_session(session: Session, metadata: dict[str, Any]) -> bool:
+    """Persist a WebUI marker only when the inbound websocket frame opted in."""
+    if metadata.get(WEBUI_SESSION_METADATA_KEY) is not True:
+        return False
+    session.metadata[WEBUI_SESSION_METADATA_KEY] = True
+    return True
+
+
+def clean_generated_title(raw: str | None) -> str:
+    text = (raw or "").strip()
+    if not text:
+        return ""
+    text = re.sub(r"^\s*(title|标题)\s*[:：]\s*", "", text, flags=re.IGNORECASE)
+    text = text.strip().strip("\"'`“”‘’")
+    text = re.sub(r"\s+", " ", text).strip()
+    text = text.rstrip("。.!！?？,，;；:")
+    if len(text) > TITLE_MAX_CHARS:
+        text = text[: TITLE_MAX_CHARS - 1].rstrip() + "…"
+    return text
+
+
+def _title_inputs(session: Session) -> tuple[str, str]:
+    user_text = ""
+    assistant_text = ""
+    for message in session.messages:
+        if message.get("_command") is True:
+            continue
+        role = message.get("role")
+        content = message.get("content")
+        if not isinstance(content, str) or not content.strip():
+            continue
+        if role == "user" and not user_text:
+            user_text = content.strip()
+        elif role == "assistant" and not assistant_text:
+            assistant_text = content.strip()
+        if user_text and assistant_text:
+            break
+    return user_text, assistant_text
+
+
+async def maybe_generate_webui_title(
+    *,
+    sessions: SessionManager,
+    session_key: str,
+    provider: LLMProvider,
+    model: str,
+) -> bool:
+    """Generate and persist a short title for WebUI-owned sessions only."""
+    session = sessions.get_or_create(session_key)
+    if session.metadata.get(WEBUI_SESSION_METADATA_KEY) is not True:
+        return False
+    if session.metadata.get(WEBUI_TITLE_USER_EDITED_METADATA_KEY) is True:
+        return False
+    current_title = session.metadata.get(WEBUI_TITLE_METADATA_KEY)
+    if isinstance(current_title, str) and current_title.strip():
+        return False
+
+    user_text, assistant_text = _title_inputs(session)
+    if not user_text:
+        return False
+
+    prompt = (
+        "Generate a concise title for this chat.\n"
+        "Rules:\n"
+        "- Use the same language as the user when practical.\n"
+        "- 3 to 8 words.\n"
+        "- No quotes.\n"
+        "- No punctuation at the end.\n"
+        "- Return only the title.\n\n"
+        f"User: {truncate_text(user_text, 1_000)}"
+    )
+    if assistant_text:
+        prompt += f"\nAssistant: {truncate_text(assistant_text, 1_000)}"
+
+    try:
+        response = await provider.chat_with_retry(
+            [
+                {
+                    "role": "system",
+                    "content": (
+                        "You write short, neutral chat titles. "
+                        "Return only the title text."
+                    ),
+                },
+                {"role": "user", "content": prompt},
+            ],
+            tools=None,
+            model=model,
+            max_tokens=TITLE_GENERATION_MAX_TOKENS,
+            temperature=0.2,
+            reasoning_effort=TITLE_GENERATION_REASONING_EFFORT,
+            retry_mode="standard",
+        )
+    except Exception:
+        logger.debug("Failed to generate webui session title for {}", session_key, exc_info=True)
+        return False
+
+    title = clean_generated_title(response.content)
+    if not title or title.lower().startswith("error"):
+        logger.debug(
+            "WebUI title generation returned no usable title for {} (finish_reason={})",
+            session_key,
+            response.finish_reason,
+        )
+        return False
+    session.metadata[WEBUI_TITLE_METADATA_KEY] = title
+    sessions.save(session)
+    return True
+
+
+async def maybe_generate_webui_title_after_turn(
+    *,
+    channel: str,
+    metadata: dict[str, Any],
+    sessions: SessionManager,
+    session_key: str,
+    provider: LLMProvider,
+    model: str,
+) -> bool:
+    if channel != "websocket" or metadata.get(WEBUI_SESSION_METADATA_KEY) is not True:
+        return False
+    return await maybe_generate_webui_title(
+        sessions=sessions,
+        session_key=session_key,
+        provider=provider,
+        model=model,
+    )
+
+
 def websocket_turn_wall_started_at(chat_id: str) -> float | None:
     """Return ``time.time()`` when the active user turn began, if still running."""
     return _WEBSOCKET_TURN_WALL_STARTED_AT.get(chat_id)
@@ -46,3 +192,156 @@ async def publish_turn_run_status(bus: MessageBus, msg: InboundMessage, status:
             metadata=meta,
         ),
     )
+
+
+def build_bus_progress_callback(
+    bus: MessageBus,
+    msg: InboundMessage,
+) -> Callable[..., Awaitable[None]]:
+    """Return the bus progress callback for agent runtime events."""
+
+    async def _publish_progress(
+        content: str,
+        *,
+        tool_hint: bool = False,
+        tool_events: list[dict[str, Any]] | None = None,
+        file_edit_events: list[dict[str, Any]] | None = None,
+        reasoning: bool = False,
+        reasoning_end: bool = False,
+    ) -> None:
+        meta = dict(msg.metadata or {})
+        meta["_progress"] = True
+        meta["_tool_hint"] = tool_hint
+        if reasoning:
+            meta["_reasoning_delta"] = True
+        if reasoning_end:
+            meta["_reasoning_end"] = True
+        if tool_events:
+            meta["_tool_events"] = tool_events
+        if file_edit_events:
+            meta["_file_edit_events"] = file_edit_events
+        await bus.publish_outbound(
+            OutboundMessage(
+                channel=msg.channel,
+                chat_id=msg.chat_id,
+                content=content,
+                metadata=meta,
+            )
+        )
+
+    if msg.channel == "websocket":
+        async def _websocket_progress(
+            content: str,
+            *,
+            tool_hint: bool = False,
+            tool_events: list[dict[str, Any]] | None = None,
+            file_edit_events: list[dict[str, Any]] | None = None,
+            reasoning: bool = False,
+            reasoning_end: bool = False,
+        ) -> None:
+            await _publish_progress(
+                content,
+                tool_hint=tool_hint,
+                tool_events=tool_events,
+                file_edit_events=file_edit_events,
+                reasoning=reasoning,
+                reasoning_end=reasoning_end,
+            )
+
+        return _websocket_progress
+
+    async def _bus_progress(
+        content: str,
+        *,
+        tool_hint: bool = False,
+        tool_events: list[dict[str, Any]] | None = None,
+        reasoning: bool = False,
+        reasoning_end: bool = False,
+    ) -> None:
+        await _publish_progress(
+            content,
+            tool_hint=tool_hint,
+            tool_events=tool_events,
+            reasoning=reasoning,
+            reasoning_end=reasoning_end,
+        )
+
+    return _bus_progress
+
+
+@dataclass
+class WebuiTurnCoordinator:
+    """Own the WebUI/WebSocket wire details that hang off AgentLoop turns."""
+
+    bus: MessageBus
+    sessions: SessionManager
+    schedule_background: Callable[[Awaitable[None]], None]
+    _title_contexts: dict[str, LLMRuntime] = field(default_factory=dict)
+
+    def capture_title_context(
+        self,
+        session_key: str,
+        msg: InboundMessage,
+        llm: LLMRuntime,
+    ) -> None:
+        if msg.channel == "websocket" and msg.metadata.get("webui") is True:
+            self._title_contexts[session_key] = llm
+
+    def discard(self, session_key: str) -> None:
+        self._title_contexts.pop(session_key, None)
+
+    async def publish_run_status(self, msg: InboundMessage, status: str) -> None:
+        await publish_turn_run_status(self.bus, msg, status)
+
+    async def handle_turn_end(
+        self,
+        msg: InboundMessage,
+        *,
+        session_key: str,
+        latency_ms: int | None,
+    ) -> None:
+        if msg.channel != "websocket":
+            return
+
+        turn_metadata: dict[str, Any] = {**msg.metadata, "_turn_end": True}
+        if latency_ms is not None:
+            turn_metadata["latency_ms"] = int(latency_ms)
+        session = self.sessions.get_or_create(session_key)
+        turn_metadata["goal_state"] = goal_state_ws_blob(session.metadata)
+        await self.bus.publish_outbound(OutboundMessage(
+            channel=msg.channel,
+            chat_id=msg.chat_id,
+            content="",
+            metadata=turn_metadata,
+        ))
+        self._schedule_title_update(msg, session_key=session_key)
+
+    def _schedule_title_update(self, msg: InboundMessage, *, session_key: str) -> None:
+        title_context = self._title_contexts.pop(session_key, None)
+        if msg.metadata.get("webui") is not True or title_context is None:
+            return
+
+        async def _generate_title_and_notify(
+            title_llm: LLMRuntime = title_context,
+        ) -> None:
+            generated = await maybe_generate_webui_title_after_turn(
+                channel=msg.channel,
+                metadata=msg.metadata,
+                sessions=self.sessions,
+                session_key=session_key,
+                provider=title_llm.provider,
+                model=title_llm.model,
+            )
+            if generated:
+                await self.bus.publish_outbound(OutboundMessage(
+                    channel=msg.channel,
+                    chat_id=msg.chat_id,
+                    content="",
+                    metadata={
+                        **msg.metadata,
+                        "_session_updated": True,
+                        "_session_update_scope": "metadata",
+                    },
+                ))
+
+        self.schedule_background(_generate_title_and_notify())

nanobot/web/__init__.py

@@ -1,6 +1,8 @@
 """Embedded web UI assets.
 
-The ``dist/`` subdirectory is populated by ``cd webui && bun run build`` and
-is shipped in the wheel; it stays empty in source checkouts until that command
-has been run.
+The ``dist/`` subdirectory holds the production WebUI bundle served by the
+gateway. It is shipped inside the published wheel and is rebuilt automatically
+by the ``webui-build`` Hatch hook during ``python -m build``. In an editable
+source checkout it stays empty until you run ``cd webui && bun run build``
+(or use the Vite dev server at ``cd webui && bun run dev``).
 """

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "nanobot-ai"
-version = "0.1.5.post3"
+version = "0.2.0"
 description = "A lightweight personal AI assistant framework"
 readme = { file = "README.md", content-type = "text/markdown" }
 requires-python = ">=3.11"
@@ -121,12 +121,22 @@ build-backend = "hatchling.build"
 [tool.hatch.metadata]
 allow-direct-references = true
 
+[tool.hatch.build.hooks.custom]
+# Implementation lives in the conventional `hatch_build.py` at the repo root.
+
 [tool.hatch.build]
 include = [
     "nanobot/**/*.py",
     "nanobot/templates/**/*.md",
     "nanobot/skills/**/*.md",
     "nanobot/skills/**/*.sh",
+    "nanobot/web/dist/**/*",
+]
+# nanobot/web/dist/ is produced by `cd webui && bun run build` and is
+# git-ignored. List it as an artifact so hatch ships it in both wheel and
+# sdist even though VCS does not track it.
+artifacts = [
+    "nanobot/web/dist/**/*",
 ]
 
 [tool.hatch.build.targets.wheel]
@@ -141,7 +151,9 @@ packages = ["nanobot"]
 [tool.hatch.build.targets.sdist]
 include = [
     "nanobot/",
+    "nanobot/web/dist/",
     "bridge/",
+    "hatch_build.py",
     "README.md",
     "LICENSE",
     "THIRD_PARTY_NOTICES.md",

tests/agent/test_auto_compact.py

@@ -45,6 +45,73 @@ def _add_turns(session, turns: int, *, prefix: str = "msg") -> None:
         session.add_message("assistant", f"{prefix} assistant {i}")
 
 
+def _make_fake_compact(
+    loop: AgentLoop,
+    *,
+    summary: str = "Summary.",
+    on_archive=None,
+    track_archived: list | None = None,
+    track_count: bool = False,
+):
+    """Return a fake compact_idle_session that mirrors the real method's session mutation."""
+    from nanobot.session.manager import Session as _Session
+
+    state = {"count": 0}
+
+    async def _fake_compact(key: str, max_suffix: int = 8) -> str:
+        state["count"] += 1
+        session = loop.sessions.get_or_create(key)
+
+        tail = list(session.messages[session.last_consolidated:])
+        if not tail:
+            session.updated_at = datetime.now()
+            loop.sessions.save(session)
+            return ""
+
+        probe = _Session(
+            key=session.key,
+            messages=tail.copy(),
+            created_at=session.created_at,
+            updated_at=session.updated_at,
+            metadata={},
+            last_consolidated=0,
+        )
+        probe.retain_recent_legal_suffix(max_suffix)
+        kept = probe.messages
+        cut = len(tail) - len(kept)
+        archive_msgs = tail[:cut]
+
+        if not archive_msgs and not kept:
+            session.updated_at = datetime.now()
+            loop.sessions.save(session)
+            return ""
+
+        last_active = session.updated_at
+        s = summary
+        if archive_msgs:
+            if on_archive:
+                result = on_archive(archive_msgs)
+                s = result if isinstance(result, str) else summary
+            if track_archived is not None:
+                track_archived.extend(archive_msgs)
+
+        if s and s != "(nothing)":
+            session.metadata["_last_summary"] = {
+                "text": s,
+                "last_active": last_active.isoformat(),
+            }
+
+        session.messages = kept
+        session.last_consolidated = 0
+        session.updated_at = datetime.now()
+        loop.sessions.save(session)
+        return s
+
+    # Attach state for count access
+    _fake_compact.state = state  # type: ignore[attr-defined]
+    return _fake_compact
+
+
 class TestSessionTTLConfig:
     """Test session TTL configuration."""
 
@@ -201,10 +268,7 @@ class TestAutoCompact:
         s2.add_message("user", "recent")
         loop.sessions.save(s2)
 
-        async def _fake_archive(messages):
-            return "Summary."
-
-        loop.consolidator.archive = _fake_archive
+        loop.consolidator.compact_idle_session = _make_fake_compact(loop)
         loop.auto_compact.check_expired(loop._schedule_background)
         await asyncio.sleep(0.1)
 
@@ -222,12 +286,9 @@ class TestAutoCompact:
         loop.sessions.save(session)
 
         archived_messages = []
-
-        async def _fake_archive(messages):
-            archived_messages.extend(messages)
-            return "Summary."
-
-        loop.consolidator.archive = _fake_archive
+        loop.consolidator.compact_idle_session = _make_fake_compact(
+            loop, track_archived=archived_messages,
+        )
 
         await loop.auto_compact._archive("cli:test")
 
@@ -246,10 +307,9 @@ class TestAutoCompact:
         _add_turns(session, 6, prefix="hello")
         loop.sessions.save(session)
 
-        async def _fake_archive(messages):
-            return "User said hello."
-
-        loop.consolidator.archive = _fake_archive
+        loop.consolidator.compact_idle_session = _make_fake_compact(
+            loop, summary="User said hello.",
+        )
 
         await loop.auto_compact._archive("cli:test")
 
@@ -262,23 +322,16 @@ class TestAutoCompact:
 
     @pytest.mark.asyncio
     async def test_auto_compact_empty_session(self, tmp_path):
-        """_archive on empty session should not archive."""
+        """_archive on empty session should not store a summary."""
         loop = _make_loop(tmp_path, session_ttl_minutes=15)
 
-        archive_called = False
-
-        async def _fake_archive(messages):
-            nonlocal archive_called
-            archive_called = True
-            return "Summary."
-
-        loop.consolidator.archive = _fake_archive
+        loop.consolidator.compact_idle_session = _make_fake_compact(loop)
 
         await loop.auto_compact._archive("cli:test")
 
-        assert not archive_called
         session_after = loop.sessions.get_or_create("cli:test")
         assert len(session_after.messages) == 0
+        assert "cli:test" not in loop.auto_compact._summaries
         await loop.close_mcp()
 
     @pytest.mark.asyncio
@@ -290,18 +343,14 @@ class TestAutoCompact:
         session.last_consolidated = 18
         loop.sessions.save(session)
 
-        archived_count = 0
-
-        async def _fake_archive(messages):
-            nonlocal archived_count
-            archived_count = len(messages)
-            return "Summary."
-
-        loop.consolidator.archive = _fake_archive
+        archived_messages = []
+        loop.consolidator.compact_idle_session = _make_fake_compact(
+            loop, track_archived=archived_messages,
+        )
 
         await loop.auto_compact._archive("cli:test")
 
-        assert archived_count == 2
+        assert len(archived_messages) == 2
         await loop.close_mcp()
 
 
@@ -334,12 +383,9 @@ class TestAutoCompactIdleDetection:
         loop.sessions.save(session)
 
         archived_messages = []
-
-        async def _fake_archive(messages):
-            archived_messages.extend(messages)
-            return "Summary."
-
-        loop.consolidator.archive = _fake_archive
+        loop.consolidator.compact_idle_session = _make_fake_compact(
+            loop, track_archived=archived_messages,
+        )
 
         # Simulate proactive archive completing before message arrives
         await loop.auto_compact._archive("cli:test")
@@ -402,10 +448,7 @@ class TestAutoCompactIdleDetection:
         session.updated_at = datetime.now() - timedelta(minutes=20)
         loop.sessions.save(session)
 
-        async def _fake_archive(messages):
-            return "Summary."
-
-        loop.consolidator.archive = _fake_archive
+        loop.consolidator.compact_idle_session = _make_fake_compact(loop)
 
         msg = InboundMessage(channel="cli", sender_id="user", chat_id="test", content="/new")
         response = await loop._process_message(msg)
@@ -466,10 +509,7 @@ class TestAutoCompactSystemMessages:
         session.updated_at = datetime.now() - timedelta(minutes=20)
         loop.sessions.save(session)
 
-        async def _fake_archive(messages):
-            return "Summary."
-
-        loop.consolidator.archive = _fake_archive
+        loop.consolidator.compact_idle_session = _make_fake_compact(loop)
 
         # Simulate proactive archive completing before system message arrives
         await loop.auto_compact._archive("cli:test")
@@ -547,12 +587,9 @@ class TestAutoCompactEdgeCases:
         loop.sessions.save(session)
 
         archived_messages = []
-
-        async def _fake_archive(messages):
-            archived_messages.extend(messages)
-            return "Summary."
-
-        loop.consolidator.archive = _fake_archive
+        loop.consolidator.compact_idle_session = _make_fake_compact(
+            loop, track_archived=archived_messages,
+        )
 
         # Simulate proactive archive completing before message arrives
         await loop.auto_compact._archive("cli:test")
@@ -644,10 +681,7 @@ class TestAutoCompactIntegration:
         session.updated_at = datetime.now() - timedelta(minutes=20)
         loop.sessions.save(session)
 
-        async def _fake_archive(messages):
-            return "Summary."
-
-        loop.consolidator.archive = _fake_archive
+        loop.consolidator.compact_idle_session = _make_fake_compact(loop)
 
         # Simulate proactive archive completing before message arrives
         await loop.auto_compact._archive("cli:test")
@@ -704,12 +738,9 @@ class TestProactiveAutoCompact:
         loop.sessions.save(session)
 
         archived_messages = []
-
-        async def _fake_archive(messages):
-            archived_messages.extend(messages)
-            return "User chatted about old things."
-
-        loop.consolidator.archive = _fake_archive
+        loop.consolidator.compact_idle_session = _make_fake_compact(
+            loop, summary="User chatted about old things.", track_archived=archived_messages,
+        )
 
         await self._run_check_expired(loop)
 
@@ -748,14 +779,14 @@ class TestProactiveAutoCompact:
         started = asyncio.Event()
         block_forever = asyncio.Event()
 
-        async def _slow_archive(messages):
+        async def _slow_compact(key, max_suffix=8):
             nonlocal archive_count
             archive_count += 1
             started.set()
             await block_forever.wait()
             return "Summary."
 
-        loop.consolidator.archive = _slow_archive
+        loop.consolidator.compact_idle_session = _slow_compact
 
         # First call starts archiving via callback
         loop.auto_compact.check_expired(loop._schedule_background)
@@ -781,10 +812,10 @@ class TestProactiveAutoCompact:
         session.updated_at = datetime.now() - timedelta(minutes=20)
         loop.sessions.save(session)
 
-        async def _failing_archive(messages):
+        async def _failing_compact(key, max_suffix=8):
             raise RuntimeError("LLM down")
 
-        loop.consolidator.archive = _failing_archive
+        loop.consolidator.compact_idle_session = _failing_compact
 
         # Should not raise
         await self._run_check_expired(loop)
@@ -795,24 +826,18 @@ class TestProactiveAutoCompact:
 
     @pytest.mark.asyncio
     async def test_proactive_archive_skips_empty_sessions(self, tmp_path):
-        """Proactive archive should not call LLM for sessions with no un-consolidated messages."""
+        """Proactive archive should not produce a summary for sessions with no messages."""
         loop = _make_loop(tmp_path, session_ttl_minutes=15)
         session = loop.sessions.get_or_create("cli:test")
         session.updated_at = datetime.now() - timedelta(minutes=20)
         loop.sessions.save(session)
 
-        archive_called = False
-
-        async def _fake_archive(messages):
-            nonlocal archive_called
-            archive_called = True
-            return "Summary."
-
-        loop.consolidator.archive = _fake_archive
+        loop.consolidator.compact_idle_session = _make_fake_compact(loop)
 
         await self._run_check_expired(loop)
 
-        assert not archive_called
+        # Empty session should not produce a summary
+        assert "cli:test" not in loop.auto_compact._summaries
         await loop.close_mcp()
 
     @pytest.mark.asyncio
@@ -824,18 +849,12 @@ class TestProactiveAutoCompact:
         session.updated_at = datetime.now() - timedelta(minutes=20)
         loop.sessions.save(session)
 
-        archive_count = 0
-
-        async def _fake_archive(messages):
-            nonlocal archive_count
-            archive_count += 1
-            return "Summary."
-
-        loop.consolidator.archive = _fake_archive
+        _fake_compact = _make_fake_compact(loop)
+        loop.consolidator.compact_idle_session = _fake_compact
 
         # Simulate an active agent task for this session
         await self._run_check_expired(loop, active_session_keys={"cli:test"})
-        assert archive_count == 0
+        assert _fake_compact.state["count"] == 0
 
         session_after = loop.sessions.get_or_create("cli:test")
         assert len(session_after.messages) == 12  # All messages preserved
@@ -851,22 +870,16 @@ class TestProactiveAutoCompact:
         session.updated_at = datetime.now() - timedelta(minutes=20)
         loop.sessions.save(session)
 
-        archive_count = 0
-
-        async def _fake_archive(messages):
-            nonlocal archive_count
-            archive_count += 1
-            return "Summary."
-
-        loop.consolidator.archive = _fake_archive
+        _fake_compact = _make_fake_compact(loop)
+        loop.consolidator.compact_idle_session = _fake_compact
 
         # First tick: active task, skip
         await self._run_check_expired(loop, active_session_keys={"cli:test"})
-        assert archive_count == 0
+        assert _fake_compact.state["count"] == 0
 
         # Second tick: task completed, should archive
         await self._run_check_expired(loop)
-        assert archive_count == 1
+        assert _fake_compact.state["count"] == 1
         await loop.close_mcp()
 
     @pytest.mark.asyncio
@@ -888,18 +901,12 @@ class TestProactiveAutoCompact:
         s3.add_message("user", "recent")
         loop.sessions.save(s3)
 
-        archive_count = 0
-
-        async def _fake_archive(messages):
-            nonlocal archive_count
-            archive_count += 1
-            return "Summary."
-
-        loop.consolidator.archive = _fake_archive
+        _fake_compact = _make_fake_compact(loop)
+        loop.consolidator.compact_idle_session = _fake_compact
 
         await self._run_check_expired(loop, active_session_keys={"cli:expired_active"})
 
-        assert archive_count == 1
+        assert _fake_compact.state["count"] == 1
         s1_after = loop.sessions.get_or_create("cli:expired_idle")
         assert len(s1_after.messages) == loop.auto_compact._RECENT_SUFFIX_MESSAGES
         s2_after = loop.sessions.get_or_create("cli:expired_active")
@@ -917,22 +924,16 @@ class TestProactiveAutoCompact:
         session.updated_at = datetime.now() - timedelta(minutes=20)
         loop.sessions.save(session)
 
-        archive_count = 0
-
-        async def _fake_archive(messages):
-            nonlocal archive_count
-            archive_count += 1
-            return "Summary."
-
-        loop.consolidator.archive = _fake_archive
+        _fake_compact = _make_fake_compact(loop)
+        loop.consolidator.compact_idle_session = _fake_compact
 
         # First tick: archives the session
         await self._run_check_expired(loop)
-        assert archive_count == 1
+        assert _fake_compact.state["count"] == 1
 
         # Second tick: should NOT re-schedule (updated_at is fresh after clear)
         await self._run_check_expired(loop)
-        assert archive_count == 1  # Still 1, not re-scheduled
+        assert _fake_compact.state["count"] == 1  # Still 1, not re-scheduled
         await loop.close_mcp()
 
     @pytest.mark.asyncio
@@ -943,22 +944,15 @@ class TestProactiveAutoCompact:
         session.updated_at = datetime.now() - timedelta(minutes=20)
         loop.sessions.save(session)
 
-        archive_count = 0
-
-        async def _fake_archive(messages):
-            nonlocal archive_count
-            archive_count += 1
-            return "Summary."
-
-        loop.consolidator.archive = _fake_archive
+        loop.consolidator.compact_idle_session = _make_fake_compact(loop)
 
         # First tick: skips (no messages), refreshes updated_at
         await self._run_check_expired(loop)
-        assert archive_count == 0
+        assert "cli:test" not in loop.auto_compact._summaries
 
         # Second tick: should NOT re-schedule because updated_at is fresh
         await self._run_check_expired(loop)
-        assert archive_count == 0
+        assert "cli:test" not in loop.auto_compact._summaries
         await loop.close_mcp()
 
     @pytest.mark.asyncio
@@ -970,18 +964,12 @@ class TestProactiveAutoCompact:
         session.updated_at = datetime.now() - timedelta(minutes=20)
         loop.sessions.save(session)
 
-        archive_count = 0
-
-        async def _fake_archive(messages):
-            nonlocal archive_count
-            archive_count += 1
-            return "Summary."
-
-        loop.consolidator.archive = _fake_archive
+        _fake_compact = _make_fake_compact(loop)
+        loop.consolidator.compact_idle_session = _fake_compact
 
         # First compact cycle
         await loop.auto_compact._archive("cli:test")
-        assert archive_count == 1
+        assert _fake_compact.state["count"] == 1
 
         # User returns, sends new messages
         msg = InboundMessage(channel="cli", sender_id="user", chat_id="test", content="second topic")
@@ -995,7 +983,7 @@ class TestProactiveAutoCompact:
 
         # Second compact cycle should succeed
         await loop.auto_compact._archive("cli:test")
-        assert archive_count == 2
+        assert _fake_compact.state["count"] == 2
         await loop.close_mcp()
 
 
@@ -1011,10 +999,9 @@ class TestSummaryPersistence:
         session.updated_at = datetime.now() - timedelta(minutes=20)
         loop.sessions.save(session)
 
-        async def _fake_archive(messages):
-            return "User said hello."
-
-        loop.consolidator.archive = _fake_archive
+        loop.consolidator.compact_idle_session = _make_fake_compact(
+            loop, summary="User said hello.",
+        )
 
         await loop.auto_compact._archive("cli:test")
 
@@ -1036,10 +1023,9 @@ class TestSummaryPersistence:
         session.updated_at = last_active
         loop.sessions.save(session)
 
-        async def _fake_archive(messages):
-            return "User said hello."
-
-        loop.consolidator.archive = _fake_archive
+        loop.consolidator.compact_idle_session = _make_fake_compact(
+            loop, summary="User said hello.",
+        )
 
         # Archive
         await loop.auto_compact._archive("cli:test")
@@ -1069,10 +1055,7 @@ class TestSummaryPersistence:
         session.updated_at = datetime.now() - timedelta(minutes=20)
         loop.sessions.save(session)
 
-        async def _fake_archive(messages):
-            return "Summary."
-
-        loop.consolidator.archive = _fake_archive
+        loop.consolidator.compact_idle_session = _make_fake_compact(loop)
 
         await loop.auto_compact._archive("cli:test")
 
@@ -1100,10 +1083,7 @@ class TestSummaryPersistence:
         session.updated_at = datetime.now() - timedelta(minutes=20)
         loop.sessions.save(session)
 
-        async def _fake_archive(messages):
-            return "Summary."
-
-        loop.consolidator.archive = _fake_archive
+        loop.consolidator.compact_idle_session = _make_fake_compact(loop)
 
         await loop.auto_compact._archive("cli:test")
 
@@ -1129,10 +1109,9 @@ class TestSummaryPersistence:
         session.updated_at = datetime.now() - timedelta(minutes=20)
         loop.sessions.save(session)
 
-        async def _fake_archive(messages):
-            return "First summary."
-
-        loop.consolidator.archive = _fake_archive
+        loop.consolidator.compact_idle_session = _make_fake_compact(
+            loop, summary="First summary.",
+        )
         await loop.auto_compact._archive("cli:test")
 
         # Consume the first summary via hot path
@@ -1148,10 +1127,9 @@ class TestSummaryPersistence:
         session.updated_at = datetime.now() - timedelta(minutes=20)
         loop.sessions.save(session)
 
-        async def _fake_archive2(messages):
-            return "Second summary."
-
-        loop.consolidator.archive = _fake_archive2
+        loop.consolidator.compact_idle_session = _make_fake_compact(
+            loop, summary="Second summary.",
+        )
         await loop.auto_compact._archive("cli:test")
 
         # The second archive writes a new summary
@@ -1173,10 +1151,9 @@ class TestSummaryPersistence:
         session.updated_at = datetime.now() - timedelta(minutes=20)
         loop.sessions.save(session)
 
-        async def _fake_archive(messages):
-            return "Old summary."
-
-        loop.consolidator.archive = _fake_archive
+        loop.consolidator.compact_idle_session = _make_fake_compact(
+            loop, summary="Old summary.",
+        )
         await loop.auto_compact._archive("cli:test")
 
         # Verify summary exists before /new

tests/agent/test_autocompact_unit.py

@@ -38,7 +38,7 @@ def _make_autocompact(
         sessions = MagicMock(spec=SessionManager)
     if consolidator is None:
         consolidator = MagicMock()
-        consolidator.archive = AsyncMock(return_value="Summary.")
+        consolidator.compact_idle_session = AsyncMock(return_value="Summary.")
     return AutoCompact(
         sessions=sessions,
         consolidator=consolidator,
@@ -178,62 +178,6 @@ class TestFormatSummary:
         assert result.startswith("Previous conversation summary (last active ")
 
 
-# ---------------------------------------------------------------------------
-# _split_unconsolidated
-# ---------------------------------------------------------------------------
-
-
-class TestSplitUnconsolidated:
-    """Test AutoCompact._split_unconsolidated splitting logic."""
-
-    def test_empty_session_returns_both_empty(self):
-        """Empty session should return ([], [])."""
-        ac = _make_autocompact()
-        session = _make_session(messages=[])
-        archive, kept = ac._split_unconsolidated(session)
-        assert archive == []
-        assert kept == []
-
-    def test_all_messages_archivable_when_more_than_suffix(self):
-        """Session with many messages should archive a prefix and keep suffix."""
-        ac = _make_autocompact()
-        msgs = [{"role": "user", "content": f"u{i}"} for i in range(20)]
-        session = _make_session(messages=msgs)
-        archive, kept = ac._split_unconsolidated(session)
-        assert len(archive) > 0
-        assert len(kept) <= AutoCompact._RECENT_SUFFIX_MESSAGES
-
-    def test_fewer_messages_than_suffix_returns_empty_archive(self):
-        """Session with fewer messages than suffix should have empty archive."""
-        ac = _make_autocompact()
-        msgs = [{"role": "user", "content": f"u{i}"} for i in range(3)]
-        session = _make_session(messages=msgs)
-        archive, kept = ac._split_unconsolidated(session)
-        assert archive == []
-        assert len(kept) == len(msgs)
-
-    def test_respects_last_consolidated_offset(self):
-        """Only messages after last_consolidated should be considered."""
-        ac = _make_autocompact()
-        msgs = [{"role": "user", "content": f"u{i}"} for i in range(20)]
-        # First 10 are already consolidated
-        session = _make_session(messages=msgs, last_consolidated=10)
-        archive, kept = ac._split_unconsolidated(session)
-        # Only the tail of 10 messages is considered for splitting
-        assert all(m["content"] in [f"u{i}" for i in range(10, 20)] for m in kept)
-        assert all(m["content"] in [f"u{i}" for i in range(10, 20)] for m in archive)
-
-    def test_retain_recent_legal_suffix_keeps_last_n(self):
-        """The kept suffix should be at most _RECENT_SUFFIX_MESSAGES long."""
-        ac = _make_autocompact()
-        # 20 user messages = 20 messages total, all after last_consolidated=0
-        msgs = [{"role": "user", "content": f"u{i}"} for i in range(20)]
-        session = _make_session(messages=msgs)
-        archive, kept = ac._split_unconsolidated(session)
-        assert len(kept) <= AutoCompact._RECENT_SUFFIX_MESSAGES
-        assert len(archive) == len(msgs) - len(kept)
-
-
 # ---------------------------------------------------------------------------
 # check_expired
 # ---------------------------------------------------------------------------
@@ -313,126 +257,71 @@ class TestCheckExpired:
 # ---------------------------------------------------------------------------
 
 
-class TestArchive:
-    """Test AutoCompact._archive async method."""
+class TestArchiveDelegates:
+    """_archive should delegate all session mutation to Consolidator."""
 
     @pytest.mark.asyncio
-    async def test_empty_session_updates_timestamp_no_archive_call(self):
-        """Empty session should refresh updated_at and not call consolidator.archive."""
+    async def test_calls_compact_idle_session(self):
         ac = _make_autocompact()
         mock_sm = MagicMock(spec=SessionManager)
-        empty_session = _make_session(messages=[])
-        mock_sm.get_or_create.return_value = empty_session
         ac.sessions = mock_sm
-        ac.consolidator.archive = AsyncMock(return_value="Summary.")
+        ac.consolidator.compact_idle_session = AsyncMock(return_value="Summary.")
 
         await ac._archive("cli:test")
 
-        ac.consolidator.archive.assert_not_called()
-        mock_sm.save.assert_called_once_with(empty_session)
-        # updated_at was refreshed
-        assert empty_session.updated_at > datetime.now() - timedelta(seconds=5)
+        ac.consolidator.compact_idle_session.assert_awaited_once_with(
+            "cli:test", ac._RECENT_SUFFIX_MESSAGES,
+        )
 
     @pytest.mark.asyncio
-    async def test_archive_returns_empty_string_no_summary_stored(self):
-        """If archive returns empty string, no summary should be stored."""
+    async def test_populates_summaries_from_metadata(self):
         ac = _make_autocompact()
         mock_sm = MagicMock(spec=SessionManager)
-        msgs = [{"role": "user", "content": f"u{i}"} for i in range(20)]
-        session = _make_session(messages=msgs)
+        session = _make_session(
+            metadata={"_last_summary": {"text": "Hello.", "last_active": "2026-05-13T10:00:00"}}
+        )
         mock_sm.get_or_create.return_value = session
         ac.sessions = mock_sm
-        ac.consolidator.archive = AsyncMock(return_value="")
+        ac.consolidator.compact_idle_session = AsyncMock(return_value="Hello.")
 
         await ac._archive("cli:test")
 
-        assert "cli:test" not in ac._summaries
-
-    @pytest.mark.asyncio
-    async def test_archive_returns_nothing_no_summary_stored(self):
-        """If archive returns '(nothing)', no summary should be stored."""
-        ac = _make_autocompact()
-        mock_sm = MagicMock(spec=SessionManager)
-        msgs = [{"role": "user", "content": f"u{i}"} for i in range(20)]
-        session = _make_session(messages=msgs)
-        mock_sm.get_or_create.return_value = session
-        ac.sessions = mock_sm
-        ac.consolidator.archive = AsyncMock(return_value="(nothing)")
-
-        await ac._archive("cli:test")
-
-        assert "cli:test" not in ac._summaries
-
-    @pytest.mark.asyncio
-    async def test_archive_exception_caught_key_removed_from_archiving(self):
-        """If archive raises, exception is caught and key removed from _archiving."""
-        ac = _make_autocompact()
-        mock_sm = MagicMock(spec=SessionManager)
-        msgs = [{"role": "user", "content": f"u{i}"} for i in range(20)]
-        session = _make_session(messages=msgs)
-        mock_sm.get_or_create.return_value = session
-        ac.sessions = mock_sm
-        ac.consolidator.archive = AsyncMock(side_effect=RuntimeError("LLM down"))
-
-        # Should not raise
-        await ac._archive("cli:test")
-
-        assert "cli:test" not in ac._archiving
-
-    @pytest.mark.asyncio
-    async def test_successful_archive_stores_summary_in_summaries_and_metadata(self):
-        """Successful archive should store summary in _summaries dict and metadata."""
-        ac = _make_autocompact()
-        mock_sm = MagicMock(spec=SessionManager)
-        msgs = [{"role": "user", "content": f"u{i}"} for i in range(20)]
-        last_active = datetime(2026, 5, 13, 10, 0, 0)
-        session = _make_session(messages=msgs, updated_at=last_active)
-        mock_sm.get_or_create.return_value = session
-        ac.sessions = mock_sm
-        ac.consolidator.archive = AsyncMock(return_value="User discussed AI.")
-
-        await ac._archive("cli:test")
-
-        # _summaries
         entry = ac._summaries.get("cli:test")
         assert entry is not None
-        assert entry[0] == "User discussed AI."
-        assert entry[1] == last_active
-        # metadata
-        meta = session.metadata.get("_last_summary")
-        assert meta is not None
-        assert meta["text"] == "User discussed AI."
-        assert "last_active" in meta
+        assert entry[0] == "Hello."
 
     @pytest.mark.asyncio
-    async def test_finally_block_always_removes_from_archiving(self):
-        """Finally block should always remove key from _archiving, even on error."""
+    async def test_no_summary_when_compact_returns_empty(self):
         ac = _make_autocompact()
         mock_sm = MagicMock(spec=SessionManager)
-        msgs = [{"role": "user", "content": f"u{i}"} for i in range(20)]
-        session = _make_session(messages=msgs)
-        mock_sm.get_or_create.return_value = session
         ac.sessions = mock_sm
-        ac.consolidator.archive = AsyncMock(side_effect=RuntimeError("fail"))
+        ac.consolidator.compact_idle_session = AsyncMock(return_value="")
 
-        # Pre-add key to archiving to verify it gets removed
-        ac._archiving.add("cli:test")
         await ac._archive("cli:test")
-        assert "cli:test" not in ac._archiving
+
+        assert "cli:test" not in ac._summaries
 
     @pytest.mark.asyncio
-    async def test_finally_removes_from_archiving_on_success(self):
-        """Finally block should remove key from _archiving on success too."""
+    async def test_no_summary_when_compact_returns_nothing(self):
         ac = _make_autocompact()
         mock_sm = MagicMock(spec=SessionManager)
-        msgs = [{"role": "user", "content": f"u{i}"} for i in range(20)]
-        session = _make_session(messages=msgs)
-        mock_sm.get_or_create.return_value = session
         ac.sessions = mock_sm
-        ac.consolidator.archive = AsyncMock(return_value="Summary.")
+        ac.consolidator.compact_idle_session = AsyncMock(return_value="(nothing)")
+
+        await ac._archive("cli:test")
+
+        assert "cli:test" not in ac._summaries
+
+    @pytest.mark.asyncio
+    async def test_exception_still_removes_from_archiving(self):
+        ac = _make_autocompact()
+        mock_sm = MagicMock(spec=SessionManager)
+        ac.sessions = mock_sm
+        ac.consolidator.compact_idle_session = AsyncMock(side_effect=RuntimeError("fail"))
 
         ac._archiving.add("cli:test")
         await ac._archive("cli:test")
+
         assert "cli:test" not in ac._archiving
 
 

tests/agent/test_consolidator.py

@@ -28,6 +28,12 @@ def mock_provider():
 def consolidator(store, mock_provider):
     sessions = MagicMock()
     sessions.save = MagicMock()
+    # When maybe_consolidate_by_tokens refreshes the session reference via
+    # get_or_create(session.key), it should get back the same object the test
+    # passed in.  Store sessions by key so the lookup is transparent.
+    _session_cache: dict[str, MagicMock] = {}
+    sessions.get_or_create = MagicMock(side_effect=lambda key: _session_cache.get(key, MagicMock()))
+    sessions._session_cache = _session_cache
     return Consolidator(
         store=store,
         provider=mock_provider,
@@ -117,6 +123,7 @@ class TestConsolidatorTokenBudget:
         session.last_consolidated = 0
         session.messages = [{"role": "user", "content": "hi"}]
         session.key = "test:key"
+        consolidator.sessions._session_cache[session.key] = session
         consolidator.estimate_session_prompt_tokens = MagicMock(return_value=(100, "tiktoken"))
         consolidator.archive = AsyncMock(return_value=True)
         await consolidator.maybe_consolidate_by_tokens(session)
@@ -152,6 +159,7 @@ class TestConsolidatorTokenBudget:
             session.add_message("user", f"u{i}")
             session.add_message("assistant", f"a{i}")
 
+        consolidator.sessions._session_cache[session.key] = session
         consolidator.estimate_session_prompt_tokens = MagicMock(return_value=(100, "tiktoken"))
         consolidator.archive = AsyncMock(return_value="old conversation summary")
 
@@ -184,6 +192,7 @@ class TestConsolidatorTokenBudget:
         session.add_message("tool", "tool result", tool_call_id="call-1", name="x")
         session.add_message("assistant", "final answer")
 
+        consolidator.sessions._session_cache[session.key] = session
         consolidator.estimate_session_prompt_tokens = MagicMock(return_value=(100, "tiktoken"))
         consolidator.archive = AsyncMock(return_value="tool turn summary")
 
@@ -210,6 +219,7 @@ class TestConsolidatorTokenBudget:
             }
             for i in range(70)
         ]
+        consolidator.sessions._session_cache[session.key] = session
         consolidator.estimate_session_prompt_tokens = MagicMock(
             side_effect=[(1200, "tiktoken"), (400, "tiktoken")]
         )
@@ -238,6 +248,7 @@ class TestConsolidatorTokenBudget:
             for i in range(70)
         ]
         session.metadata = {}
+        consolidator.sessions._session_cache[session.key] = session
         consolidator.estimate_session_prompt_tokens = MagicMock(
             side_effect=[(1200, "tiktoken"), (400, "tiktoken")]
         )
@@ -263,6 +274,7 @@ class TestConsolidatorTokenBudget:
             for i in range(70)
         ]
         session.metadata = {}
+        consolidator.sessions._session_cache[session.key] = session
         # Keep estimates high so the loop would otherwise run multiple rounds.
         consolidator.estimate_session_prompt_tokens = MagicMock(
             return_value=(1200, "tiktoken")
@@ -287,6 +299,7 @@ class TestConsolidatorTokenBudget:
             }
             for i in range(70)
         ]
+        consolidator.sessions._session_cache[session.key] = session
         consolidator.estimate_session_prompt_tokens = MagicMock(
             side_effect=[(1200, "tiktoken"), (400, "tiktoken")]
         )
@@ -299,6 +312,260 @@ class TestConsolidatorTokenBudget:
         assert session.last_consolidated == 61
 
 
+class TestCompactIdleSession:
+    """Tests for Consolidator.compact_idle_session — lock-protected idle truncation."""
+
+    @pytest.fixture
+    def real_consolidator(self, store, mock_provider):
+        """Create a Consolidator with a real SessionManager (not a mock)."""
+        from nanobot.session.manager import SessionManager
+
+        sessions = SessionManager(store.workspace)
+        return Consolidator(
+            store=store,
+            provider=mock_provider,
+            model="test-model",
+            sessions=sessions,
+            context_window_tokens=1000,
+            build_messages=MagicMock(return_value=[]),
+            get_tool_definitions=MagicMock(return_value=[]),
+            max_completion_tokens=100,
+        )
+
+    @pytest.mark.asyncio
+    async def test_archives_prefix_keeps_suffix(self, real_consolidator, mock_provider):
+        """20 user/assistant turns → compact with max_suffix=8 → messages ≤ 8,
+        last_consolidated=0, _last_summary stored."""
+        mock_provider.chat_with_retry.return_value = MagicMock(
+            content="Summary of old conversation.", finish_reason="stop"
+        )
+        sessions = real_consolidator.sessions
+        session = sessions.get_or_create("cli:test")
+        for i in range(20):
+            session.add_message("user", f"user msg {i}")
+            session.add_message("assistant", f"assistant msg {i}")
+        sessions.save(session)
+
+        result = await real_consolidator.compact_idle_session("cli:test", max_suffix=8)
+        assert result == "Summary of old conversation."
+
+        reloaded = sessions.get_or_create("cli:test")
+        assert len(reloaded.messages) <= 8
+        assert reloaded.last_consolidated == 0
+        meta = reloaded.metadata.get("_last_summary")
+        assert meta is not None
+        assert meta["text"] == "Summary of old conversation."
+        assert "last_active" in meta
+
+    @pytest.mark.asyncio
+    async def test_empty_session_refreshes_timestamp(self, real_consolidator):
+        """Empty session with old updated_at → refreshed after call, returns ''."""
+        from datetime import datetime, timedelta
+
+        sessions = real_consolidator.sessions
+        session = sessions.get_or_create("cli:empty")
+        old_ts = datetime.now() - timedelta(hours=2)
+        session.updated_at = old_ts
+        sessions.save(session)
+
+        result = await real_consolidator.compact_idle_session("cli:empty")
+        assert result == ""
+
+        reloaded = sessions.get_or_create("cli:empty")
+        assert reloaded.updated_at > old_ts
+
+    @pytest.mark.asyncio
+    async def test_nothing_summary_not_stored(self, real_consolidator, mock_provider):
+        """LLM returns '(nothing)' → _last_summary NOT in metadata."""
+        mock_provider.chat_with_retry.return_value = MagicMock(
+            content="(nothing)", finish_reason="stop"
+        )
+        sessions = real_consolidator.sessions
+        session = sessions.get_or_create("cli:nothing")
+        for i in range(10):
+            session.add_message("user", f"u{i}")
+            session.add_message("assistant", f"a{i}")
+        sessions.save(session)
+
+        result = await real_consolidator.compact_idle_session("cli:nothing", max_suffix=4)
+        assert result == "(nothing)"
+
+        reloaded = sessions.get_or_create("cli:nothing")
+        assert "_last_summary" not in reloaded.metadata
+
+    @pytest.mark.asyncio
+    async def test_llm_failure_still_truncates(self, real_consolidator, mock_provider, store):
+        """LLM raises RuntimeError → raw_archive fires, session still truncated, returns None."""
+        mock_provider.chat_with_retry.side_effect = RuntimeError("LLM unavailable")
+        sessions = real_consolidator.sessions
+        session = sessions.get_or_create("cli:fail")
+        for i in range(10):
+            session.add_message("user", f"u{i}")
+            session.add_message("assistant", f"a{i}")
+        sessions.save(session)
+
+        result = await real_consolidator.compact_idle_session("cli:fail", max_suffix=4)
+        assert result is None
+
+        # raw_archive should have been called (history.jsonl gets an entry)
+        entries = store.read_unprocessed_history(since_cursor=0)
+        assert any("[RAW]" in e["content"] for e in entries)
+
+        # Session should still be truncated
+        reloaded = sessions.get_or_create("cli:fail")
+        assert len(reloaded.messages) <= 4
+
+    @pytest.mark.asyncio
+    async def test_respects_last_consolidated(self, real_consolidator, mock_provider):
+        """30 turns with last_consolidated=50 → only unconsolidated tail considered."""
+        mock_provider.chat_with_retry.return_value = MagicMock(
+            content="Tail summary.", finish_reason="stop"
+        )
+        sessions = real_consolidator.sessions
+        session = sessions.get_or_create("cli:offset")
+        for i in range(30):
+            session.add_message("user", f"u{i}")
+            session.add_message("assistant", f"a{i}")
+        session.last_consolidated = 50  # Only 10 messages unconsolidated
+        sessions.save(session)
+
+        result = await real_consolidator.compact_idle_session("cli:offset", max_suffix=4)
+        assert result == "Tail summary."
+
+        # Verify only the unconsolidated tail was processed:
+        # 10 unconsolidated messages (50-59), keep suffix of 4 → archive 6
+        archived_call = mock_provider.chat_with_retry.call_args
+        user_content = archived_call.kwargs["messages"][1]["content"]
+        # Should contain only tail messages, not early ones
+        assert "u0" not in user_content
+        assert "u25" in user_content or "a25" in user_content
+
+    @pytest.mark.asyncio
+    async def test_acquires_consolidation_lock(self, real_consolidator, mock_provider):
+        """Verify lock is held during execution."""
+        import asyncio
+
+        # Use a slow LLM response to ensure the lock is held while we check
+        started = asyncio.Event()
+
+        async def slow_chat(**kwargs):
+            started.set()
+            await asyncio.sleep(0.1)
+            return MagicMock(content="Summary.", finish_reason="stop")
+
+        mock_provider.chat_with_retry = slow_chat
+
+        sessions = real_consolidator.sessions
+        session = sessions.get_or_create("cli:lock")
+        for i in range(10):
+            session.add_message("user", f"u{i}")
+            session.add_message("assistant", f"a{i}")
+        sessions.save(session)
+
+        lock = real_consolidator.get_lock("cli:lock")
+        assert not lock.locked()
+
+        task = asyncio.ensure_future(
+            real_consolidator.compact_idle_session("cli:lock", max_suffix=4)
+        )
+        await started.wait()
+        assert lock.locked()
+        await task
+        assert not lock.locked()
+
+
+class TestConsolidatorSessionRefresh:
+    """Background consolidation must detect stale session references."""
+
+    @pytest.mark.asyncio
+    async def test_reloads_before_empty_session_guard(self, tmp_path):
+        """A stale empty reference must not skip a non-empty cached session."""
+        from nanobot.agent.memory import Consolidator, MemoryStore
+        from nanobot.session.manager import Session, SessionManager
+
+        store = MemoryStore(tmp_path)
+        provider = MagicMock()
+        provider.chat_with_retry = AsyncMock(
+            return_value=MagicMock(content="summary", finish_reason="stop")
+        )
+        provider.generation.max_tokens = 4096
+        provider.estimate_prompt_tokens = MagicMock(return_value=(10, "test"))
+        sessions = SessionManager(tmp_path)
+        consolidator = Consolidator(
+            store=store,
+            provider=provider,
+            model="test-model",
+            sessions=sessions,
+            context_window_tokens=128_000,
+            build_messages=MagicMock(return_value=[]),
+            get_tool_definitions=MagicMock(return_value=[]),
+        )
+
+        fresh = sessions.get_or_create("cli:test")
+        fresh.add_message("user", "fresh message")
+        sessions.save(fresh)
+        stale_empty = Session(key="cli:test")
+
+        seen: dict[str, Session] = {}
+
+        def estimate(session: Session):
+            seen["session"] = session
+            return 10, "test"
+
+        consolidator.estimate_session_prompt_tokens = MagicMock(side_effect=estimate)
+
+        await consolidator.maybe_consolidate_by_tokens(stale_empty)
+
+        assert seen["session"] is fresh
+
+    @pytest.mark.asyncio
+    async def test_reloads_stale_session_after_compact(self, tmp_path):
+        """After compact_idle_session replaces the session, a concurrent
+        maybe_consolidate_by_tokens with the old reference should use the
+        fresh session from cache instead of overwriting."""
+        from nanobot.agent.memory import Consolidator, MemoryStore
+        from nanobot.session.manager import SessionManager
+
+        store = MemoryStore(tmp_path)
+        provider = MagicMock()
+        provider.chat_with_retry = AsyncMock(
+            return_value=MagicMock(content="summary", finish_reason="stop")
+        )
+        provider.generation.max_tokens = 4096
+        provider.estimate_prompt_tokens = MagicMock(return_value=(10, "test"))
+        sessions = SessionManager(tmp_path)
+        consolidator = Consolidator(
+            store=store,
+            provider=provider,
+            model="test-model",
+            sessions=sessions,
+            context_window_tokens=128_000,
+            build_messages=MagicMock(return_value=[]),
+            get_tool_definitions=MagicMock(return_value=[]),
+        )
+
+        # Populate session with many messages
+        session = sessions.get_or_create("cli:test")
+        for i in range(20):
+            session.add_message("user", f"u{i}")
+            session.add_message("assistant", f"a{i}")
+        sessions.save(session)
+
+        # Simulate: background consolidation captures old reference
+        old_ref = session
+
+        # AutoCompact runs first and truncates to 8
+        await consolidator.compact_idle_session("cli:test", max_suffix=8)
+
+        # Background consolidation runs with stale reference —
+        # should detect the session was replaced and not undo the compact.
+        await consolidator.maybe_consolidate_by_tokens(old_ref)
+
+        session_after = sessions.get_or_create("cli:test")
+        # Messages should still be truncated (not restored to 40)
+        assert len(session_after.messages) <= 8
+
+
 class TestRawArchiveTruncation:
     """raw_archive() must cap entry size to avoid bloating history.jsonl."""
 

tests/agent/test_heartbeat_service.py

@@ -4,6 +4,7 @@ import pytest
 
 from nanobot.heartbeat.service import HeartbeatService
 from nanobot.providers.base import LLMProvider, LLMResponse, ToolCallRequest
+from nanobot.utils.llm_runtime import LLMRuntime
 
 
 class DummyProvider(LLMProvider):
@@ -11,9 +12,11 @@ class DummyProvider(LLMProvider):
         super().__init__()
         self._responses = list(responses)
         self.calls = 0
+        self.models: list[str | None] = []
 
     async def chat(self, *args, **kwargs) -> LLMResponse:
         self.calls += 1
+        self.models.append(kwargs.get("model"))
         if self._responses:
             return self._responses.pop(0)
         return LLMResponse(content="", tool_calls=[])
@@ -215,6 +218,51 @@ async def test_tick_suppresses_when_evaluator_says_no(tmp_path, monkeypatch) ->
     assert notified == []
 
 
+def test_tick_uses_runtime_provider_and_model(tmp_path, monkeypatch) -> None:
+    """Preset changes must apply to heartbeat decision and post-run evaluation."""
+    (tmp_path / "HEARTBEAT.md").write_text("- [ ] check runtime model", encoding="utf-8")
+
+    runtime_provider = DummyProvider([
+        LLMResponse(
+            content="",
+            tool_calls=[
+                ToolCallRequest(
+                    id="hb_1",
+                    name="heartbeat",
+                    arguments={"action": "run", "tasks": "check runtime model"},
+                )
+            ],
+        ),
+    ])
+    runtime_model = "openai/gpt-4.1"
+
+    executed: list[str] = []
+    evaluated: list[tuple[LLMProvider, str]] = []
+
+    async def _on_execute(tasks: str) -> str:
+        executed.append(tasks)
+        return "runtime model produced a user-facing update"
+
+    async def _eval_capture(response, tasks, provider, model):
+        evaluated.append((provider, model))
+        return False
+
+    service = HeartbeatService(
+        workspace=tmp_path,
+        llm_runtime=lambda: LLMRuntime(runtime_provider, runtime_model),
+        on_execute=_on_execute,
+    )
+
+    monkeypatch.setattr("nanobot.utils.evaluator.evaluate_response", _eval_capture)
+
+    asyncio.run(service._tick())
+
+    assert runtime_provider.calls == 1
+    assert runtime_provider.models == [runtime_model]
+    assert executed == ["check runtime model"]
+    assert evaluated == [(runtime_provider, runtime_model)]
+
+
 @pytest.mark.asyncio
 async def test_decide_retries_transient_error_then_succeeds(tmp_path, monkeypatch) -> None:
     provider = DummyProvider([
@@ -286,4 +334,3 @@ async def test_decide_prompt_includes_current_time(tmp_path) -> None:
     user_msg = captured_messages[1]
     assert user_msg["role"] == "user"
     assert "Current Time:" in user_msg["content"]
-

tests/agent/test_loop_progress.py

@@ -6,10 +6,15 @@ from unittest.mock import AsyncMock, MagicMock
 
 import pytest
 
+import nanobot.agent.runner as runner_module
 from nanobot.agent.loop import AgentLoop
 from nanobot.bus.events import InboundMessage
 from nanobot.bus.queue import MessageBus
 from nanobot.providers.base import LLMResponse, ToolCallRequest
+from nanobot.utils.progress_events import (
+    invoke_file_edit_progress,
+    on_progress_accepts_file_edit_events,
+)
 
 
 def _make_loop(tmp_path: Path) -> AgentLoop:
@@ -82,6 +87,142 @@ class TestToolEventProgress:
             ),
         ]
 
+    @pytest.mark.asyncio
+    async def test_write_file_emits_file_edit_progress(self, tmp_path: Path) -> None:
+        loop = _make_loop(tmp_path)
+        target = tmp_path / "foo.txt"
+        target.write_text("old\n", encoding="utf-8")
+        tool_call = ToolCallRequest(
+            id="call-write",
+            name="write_file",
+            arguments={"path": "foo.txt", "content": "new\nextra\n"},
+        )
+        calls = iter([
+            LLMResponse(content="", tool_calls=[tool_call]),
+            LLMResponse(content="Done", tool_calls=[]),
+        ])
+        loop.provider.chat_with_retry = AsyncMock(side_effect=lambda *a, **kw: next(calls))
+        loop.tools.get_definitions = MagicMock(return_value=[])
+        loop.tools.prepare_call = MagicMock(
+            return_value=(None, {"path": "foo.txt", "content": "new\nextra\n"}, None),
+        )
+
+        async def execute(name: str, params: dict) -> str:
+            target.write_text(params["content"], encoding="utf-8")
+            return "ok"
+
+        loop.tools.execute = AsyncMock(side_effect=execute)
+        file_events: list[dict] = []
+
+        async def on_progress(
+            content: str,
+            *,
+            tool_hint: bool = False,
+            tool_events: list[dict] | None = None,
+            file_edit_events: list[dict] | None = None,
+        ) -> None:
+            if file_edit_events:
+                file_events.extend(file_edit_events)
+
+        final_content, _, _, _, _ = await loop._run_agent_loop([], on_progress=on_progress)
+
+        assert final_content == "Done"
+        assert [event["phase"] for event in file_events] == ["start", "end"]
+        assert file_events[0] == {
+            "version": 1,
+            "call_id": "call-write",
+            "tool": "write_file",
+            "path": "foo.txt",
+            "phase": "start",
+            "added": 2,
+            "deleted": 1,
+            "approximate": True,
+            "status": "editing",
+        }
+        assert file_events[1]["status"] == "done"
+        assert file_events[1]["approximate"] is False
+        assert (file_events[1]["added"], file_events[1]["deleted"]) == (2, 1)
+
+    @pytest.mark.asyncio
+    async def test_file_edit_snapshot_skipped_when_progress_callback_cannot_emit_file_edits(
+        self,
+        tmp_path: Path,
+        monkeypatch: pytest.MonkeyPatch,
+    ) -> None:
+        loop = _make_loop(tmp_path)
+        target = tmp_path / "foo.txt"
+        target.write_text("old\n", encoding="utf-8")
+        tool_call = ToolCallRequest(
+            id="call-write",
+            name="write_file",
+            arguments={"path": "foo.txt", "content": "new\n"},
+        )
+        calls = iter([
+            LLMResponse(content="", tool_calls=[tool_call]),
+            LLMResponse(content="Done", tool_calls=[]),
+        ])
+        loop.provider.chat_with_retry = AsyncMock(side_effect=lambda *a, **kw: next(calls))
+        loop.tools.get_definitions = MagicMock(return_value=[])
+        loop.tools.prepare_call = MagicMock(
+            return_value=(None, {"path": "foo.txt", "content": "new\n"}, None),
+        )
+
+        async def execute(name: str, params: dict) -> str:
+            target.write_text(params["content"], encoding="utf-8")
+            return "ok"
+
+        loop.tools.execute = AsyncMock(side_effect=execute)
+        prepare_tracker = MagicMock(side_effect=AssertionError("unexpected file snapshot"))
+        monkeypatch.setattr(runner_module, "prepare_file_edit_tracker", prepare_tracker)
+
+        async def on_progress(
+            content: str,
+            *,
+            tool_hint: bool = False,
+            tool_events: list[dict] | None = None,
+        ) -> None:
+            pass
+
+        final_content, _, _, _, _ = await loop._run_agent_loop([], on_progress=on_progress)
+
+        assert final_content == "Done"
+        assert target.read_text(encoding="utf-8") == "new\n"
+        prepare_tracker.assert_not_called()
+
+    @pytest.mark.asyncio
+    async def test_exec_does_not_emit_file_edit_progress(self, tmp_path: Path) -> None:
+        loop = _make_loop(tmp_path)
+        tool_call = ToolCallRequest(
+            id="call-exec",
+            name="exec",
+            arguments={"command": "printf hi > foo.txt"},
+        )
+        calls = iter([
+            LLMResponse(content="", tool_calls=[tool_call]),
+            LLMResponse(content="Done", tool_calls=[]),
+        ])
+        loop.provider.chat_with_retry = AsyncMock(side_effect=lambda *a, **kw: next(calls))
+        loop.tools.get_definitions = MagicMock(return_value=[])
+        loop.tools.prepare_call = MagicMock(
+            return_value=(None, {"command": "printf hi > foo.txt"}, None),
+        )
+        loop.tools.execute = AsyncMock(return_value="ok")
+        file_events: list[dict] = []
+
+        async def on_progress(
+            content: str,
+            *,
+            tool_hint: bool = False,
+            tool_events: list[dict] | None = None,
+            file_edit_events: list[dict] | None = None,
+        ) -> None:
+            if file_edit_events:
+                file_events.extend(file_edit_events)
+
+        await loop._run_agent_loop([], on_progress=on_progress)
+
+        assert file_events == []
+
     @pytest.mark.asyncio
     async def test_bus_progress_forwards_tool_events_to_outbound_metadata(self, tmp_path: Path) -> None:
         """When run() handles a bus message, _tool_events lands in OutboundMessage metadata."""
@@ -130,6 +271,44 @@ class TestToolEventProgress:
         assert finish["phase"] == "end"
         assert finish["result"] == "file.txt"
 
+    @pytest.mark.asyncio
+    async def test_bus_progress_forwards_file_edit_events_for_websocket_only(self, tmp_path: Path) -> None:
+        bus = MessageBus()
+        provider = MagicMock()
+        provider.get_default_model.return_value = "test-model"
+        loop = AgentLoop(bus=bus, provider=provider, workspace=tmp_path, model="test-model")
+        edit_events = [{
+            "call_id": "call-write",
+            "tool": "write_file",
+            "path": "foo.txt",
+            "phase": "start",
+            "added": 1,
+            "deleted": 0,
+            "approximate": True,
+            "status": "editing",
+        }]
+
+        websocket_progress = await loop._build_bus_progress_callback(InboundMessage(
+            channel="websocket",
+            sender_id="u1",
+            chat_id="chat1",
+            content="edit",
+        ))
+        assert on_progress_accepts_file_edit_events(websocket_progress) is True
+        await websocket_progress("", file_edit_events=edit_events)
+        outbound = await bus.consume_outbound()
+        assert outbound.metadata["_file_edit_events"] == edit_events
+
+        telegram_progress = await loop._build_bus_progress_callback(InboundMessage(
+            channel="telegram",
+            sender_id="u1",
+            chat_id="chat2",
+            content="edit",
+        ))
+        assert on_progress_accepts_file_edit_events(telegram_progress) is False
+        await invoke_file_edit_progress(telegram_progress, edit_events)
+        assert bus.outbound_size == 0
+
     @pytest.mark.asyncio
     async def test_non_streaming_channel_does_not_publish_codex_progress_deltas(
         self,
@@ -353,8 +532,93 @@ class TestToolEventProgress:
         assert session_updated is not None
 
         assert (session_updated.metadata or {}).get("_session_updated") is True
+        assert (session_updated.metadata or {}).get("_session_update_scope") == "metadata"
         assert provider.chat_with_retry.await_count == 2
 
+    @pytest.mark.asyncio
+    async def test_webui_title_generation_uses_turn_model_snapshot(
+        self,
+        tmp_path: Path,
+        monkeypatch: pytest.MonkeyPatch,
+    ) -> None:
+        bus = MessageBus()
+        provider = MagicMock()
+        provider.get_default_model.return_value = "test-model"
+        provider.chat_with_retry = AsyncMock(return_value=LLMResponse(content="Done", tool_calls=[]))
+        loop = AgentLoop(bus=bus, provider=provider, workspace=tmp_path, model="test-model")
+        loop.tools.get_definitions = MagicMock(return_value=[])
+        loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
+
+        captured: dict[str, object] = {}
+
+        async def fake_title_after_turn(**kwargs: object) -> bool:
+            captured.update(kwargs)
+            return False
+
+        monkeypatch.setattr(
+            "nanobot.utils.webui_turn_helpers.maybe_generate_webui_title_after_turn",
+            fake_title_after_turn,
+        )
+        scheduled_title: list[object] = []
+
+        def schedule_background(coro: object) -> None:
+            name = getattr(coro, "__qualname__", "")
+            if "_generate_title_and_notify" in name:
+                scheduled_title.append(coro)
+            elif hasattr(coro, "close"):
+                coro.close()
+
+        loop._schedule_background = schedule_background  # type: ignore[method-assign]
+
+        await loop._dispatch(InboundMessage(
+            channel="websocket",
+            sender_id="u1",
+            chat_id="chat1",
+            content="say hello",
+            metadata={"webui": True},
+        ))
+
+        assert len(scheduled_title) == 1
+        loop.provider = MagicMock()
+        loop.model = "switched-after-turn"
+
+        await scheduled_title[0]  # type: ignore[misc]
+
+        assert captured["provider"] is provider
+        assert captured["model"] == "test-model"
+
+    @pytest.mark.asyncio
+    async def test_webui_command_turn_does_not_schedule_title_generation(
+        self,
+        tmp_path: Path,
+        monkeypatch: pytest.MonkeyPatch,
+    ) -> None:
+        bus = MessageBus()
+        provider = MagicMock()
+        provider.get_default_model.return_value = "test-model"
+        provider.chat_with_retry = AsyncMock(return_value=LLMResponse(content="Done", tool_calls=[]))
+        loop = AgentLoop(bus=bus, provider=provider, workspace=tmp_path, model="test-model")
+
+        async def fake_title_after_turn(**_kwargs: object) -> bool:
+            raise AssertionError("command-only turns should not generate titles")
+
+        monkeypatch.setattr(
+            "nanobot.utils.webui_turn_helpers.maybe_generate_webui_title_after_turn",
+            fake_title_after_turn,
+        )
+        scheduled: list[object] = []
+        loop._schedule_background = scheduled.append  # type: ignore[method-assign]
+
+        await loop._dispatch(InboundMessage(
+            channel="websocket",
+            sender_id="u1",
+            chat_id="chat1",
+            content="/model",
+            metadata={"webui": True},
+        ))
+
+        assert scheduled == []
+
     @pytest.mark.asyncio
     async def test_non_websocket_dispatch_does_not_publish_turn_end_marker(self, tmp_path: Path) -> None:
         bus = MessageBus()

tests/agent/test_loop_save_turn.py

@@ -10,12 +10,16 @@ from nanobot.bus.events import InboundMessage
 from nanobot.bus.queue import MessageBus
 from nanobot.providers.base import LLMResponse
 from nanobot.session.goal_state import GOAL_STATE_KEY
-from nanobot.session.manager import Session
-from nanobot.utils.webui_titles import (
+from nanobot.session.manager import Session, SessionManager
+from nanobot.utils.webui_turn_helpers import (
+    TITLE_GENERATION_MAX_TOKENS,
+    TITLE_GENERATION_REASONING_EFFORT,
     WEBUI_SESSION_METADATA_KEY,
     WEBUI_TITLE_METADATA_KEY,
+    WebuiTurnCoordinator,
     maybe_generate_webui_title,
 )
+from nanobot.utils.llm_runtime import LLMRuntime
 
 
 def _mk_loop() -> AgentLoop:
@@ -33,6 +37,22 @@ def _make_full_loop(tmp_path: Path) -> AgentLoop:
     return AgentLoop(bus=MessageBus(), provider=provider, workspace=tmp_path, model="test-model")
 
 
+def test_agent_loop_llm_runtime_reflects_current_provider_and_model(tmp_path: Path) -> None:
+    loop = _make_full_loop(tmp_path)
+    runtime = loop.llm_runtime()
+
+    assert runtime.provider is loop.provider
+    assert runtime.model == "test-model"
+
+    next_provider = MagicMock()
+    loop.provider = next_provider
+    loop.model = "next-model"
+    runtime = loop.llm_runtime()
+
+    assert runtime.provider is next_provider
+    assert runtime.model == "next-model"
+
+
 @pytest.mark.asyncio
 async def test_generate_webui_title_only_for_marked_webui_sessions(tmp_path: Path) -> None:
     loop = _make_full_loop(tmp_path)
@@ -55,6 +75,11 @@ async def test_generate_webui_title_only_for_marked_webui_sessions(tmp_path: Pat
     assert generated is True
     assert session.metadata[WEBUI_TITLE_METADATA_KEY] == "优化 WebUI 侧边栏"
     loop.provider.chat_with_retry.assert_awaited_once()
+    assert loop.provider.chat_with_retry.await_args.kwargs["max_tokens"] == TITLE_GENERATION_MAX_TOKENS
+    assert (
+        loop.provider.chat_with_retry.await_args.kwargs["reasoning_effort"]
+        == TITLE_GENERATION_REASONING_EFFORT
+    )
 
 
 @pytest.mark.asyncio
@@ -79,6 +104,80 @@ async def test_generate_webui_title_skips_plain_websocket_sessions(tmp_path: Pat
     loop.provider.chat_with_retry.assert_not_awaited()
 
 
+@pytest.mark.asyncio
+async def test_generate_webui_title_ignores_command_only_sessions(tmp_path: Path) -> None:
+    loop = _make_full_loop(tmp_path)
+    session = loop.sessions.get_or_create("websocket:command-title")
+    session.metadata[WEBUI_SESSION_METADATA_KEY] = True
+    session.add_message("user", "/model deep", _command=True)
+    session.add_message(
+        "assistant",
+        "Switched model preset to `deep`.\n- Model: `deepseek-v4-pro`",
+        _command=True,
+    )
+    loop.sessions.save(session)
+
+    generated = await maybe_generate_webui_title(
+        sessions=loop.sessions,
+        session_key="websocket:command-title",
+        provider=loop.provider,
+        model=loop.model,
+    )
+
+    assert generated is False
+    assert WEBUI_TITLE_METADATA_KEY not in session.metadata
+    loop.provider.chat_with_retry.assert_not_awaited()
+
+
+def test_webui_title_update_uses_captured_llm_runtime(
+    tmp_path: Path,
+    monkeypatch: pytest.MonkeyPatch,
+) -> None:
+    bus = MessageBus()
+    sessions = SessionManager(tmp_path)
+    scheduled: list[object] = []
+    captured: dict[str, object] = {}
+
+    async def fake_title_after_turn(**kwargs: object) -> bool:
+        captured.update(kwargs)
+        return False
+
+    monkeypatch.setattr(
+        "nanobot.utils.webui_turn_helpers.maybe_generate_webui_title_after_turn",
+        fake_title_after_turn,
+    )
+    coordinator = WebuiTurnCoordinator(
+        bus=bus,
+        sessions=sessions,
+        schedule_background=lambda coro: scheduled.append(coro),
+    )
+    provider = MagicMock()
+    msg = InboundMessage(
+        channel="websocket",
+        sender_id="u1",
+        chat_id="chat1",
+        content="say hello",
+        metadata={"webui": True},
+    )
+
+    coordinator.capture_title_context(
+        "websocket:chat1",
+        msg,
+        LLMRuntime(provider, "turn-model"),
+    )
+    asyncio.run(coordinator.handle_turn_end(
+        msg,
+        session_key="websocket:chat1",
+        latency_ms=None,
+    ))
+
+    assert len(scheduled) == 1
+    asyncio.run(scheduled[0])  # type: ignore[arg-type]
+
+    assert captured["provider"] is provider
+    assert captured["model"] == "turn-model"
+
+
 def test_save_turn_skips_multimodal_user_when_only_runtime_context() -> None:
     loop = _mk_loop()
     session = Session(key="test:runtime-only")

tests/agent/test_onboard_logic.py

@@ -1074,3 +1074,242 @@ class TestConfigurePydanticModelEmptyString:
         result = _configure_pydantic_model(model, "Test")
         assert result is not None
         assert result.api_key == ""
+
+
+class TestModelPresetWizard:
+    """Tests for model preset CRUD in the onboard wizard."""
+
+    def test_sync_preset_cache(self):
+        """_sync_preset_cache should populate the module-level cache."""
+        from nanobot.cli.onboard import _MODEL_PRESET_CACHE, _sync_preset_cache
+        from nanobot.config.schema import ModelPresetConfig
+
+        config = Config()
+        config.model_presets["fast"] = ModelPresetConfig(model="gpt-4.1-mini")
+        config.model_presets["power"] = ModelPresetConfig(model="gpt-4.1")
+        _sync_preset_cache(config)
+        assert _MODEL_PRESET_CACHE == {"fast", "power"}
+        _MODEL_PRESET_CACHE.clear()
+
+    def test_model_preset_add(self, monkeypatch):
+        """_configure_model_presets should add a new preset."""
+        from nanobot.cli.onboard import _MODEL_PRESET_CACHE, _configure_model_presets
+        from nanobot.config.schema import ModelPresetConfig
+
+        config = Config()
+        _MODEL_PRESET_CACHE.clear()
+
+        responses = iter([
+            "[+] Add new preset",
+            "my-preset",
+            "<- Back",
+        ])
+
+        class FakePrompt:
+            def __init__(self, response):
+                self.response = response
+
+            def ask(self):
+                if isinstance(self.response, BaseException):
+                    raise self.response
+                return self.response
+
+        def fake_select(*_args, **_kwargs):
+            return FakePrompt(next(responses))
+
+        def fake_text(*_args, **_kwargs):
+            return FakePrompt(next(responses))
+
+        def fake_configure(*_model, **_kwargs):
+            return ModelPresetConfig(model="gpt-test", temperature=0.5)
+
+        def fake_select_with_back(*_args, **_kwargs):
+            return next(responses)
+
+        monkeypatch.setattr(onboard_wizard, "_select_with_back", fake_select_with_back)
+        monkeypatch.setattr(
+            onboard_wizard, "questionary", SimpleNamespace(select=fake_select, text=fake_text)
+        )
+        monkeypatch.setattr(onboard_wizard, "_configure_pydantic_model", fake_configure)
+        monkeypatch.setattr(onboard_wizard, "_show_section_header", lambda *a, **kw: None)
+        monkeypatch.setattr(onboard_wizard, "console", SimpleNamespace(clear=lambda: None))
+
+        _configure_model_presets(config)
+
+        assert "my-preset" in config.model_presets
+        assert config.model_presets["my-preset"].model == "gpt-test"
+        assert config.model_presets["my-preset"].temperature == 0.5
+        _MODEL_PRESET_CACHE.clear()
+
+    def test_model_preset_delete(self, monkeypatch):
+        """_configure_model_presets should delete an existing preset."""
+        from nanobot.cli.onboard import _MODEL_PRESET_CACHE, _configure_model_presets
+        from nanobot.config.schema import ModelPresetConfig
+
+        config = Config()
+        config.model_presets["old"] = ModelPresetConfig(model="x")
+        _MODEL_PRESET_CACHE.clear()
+        _MODEL_PRESET_CACHE.update({"old", "default"})
+
+        responses = iter([
+            "old (x)",
+            "Delete",
+            True,
+            "<- Back",
+        ])
+
+        class FakePrompt:
+            def __init__(self, response):
+                self.response = response
+
+            def ask(self):
+                if isinstance(self.response, BaseException):
+                    raise self.response
+                return self.response
+
+        def fake_select(*_args, **_kwargs):
+            return FakePrompt(next(responses))
+
+        def fake_confirm(*_args, **_kwargs):
+            return FakePrompt(next(responses))
+
+        def fake_select_with_back(*_args, **_kwargs):
+            return next(responses)
+
+        monkeypatch.setattr(onboard_wizard, "_select_with_back", fake_select_with_back)
+        monkeypatch.setattr(
+            onboard_wizard, "questionary", SimpleNamespace(select=fake_select, confirm=fake_confirm)
+        )
+        monkeypatch.setattr(onboard_wizard, "_show_section_header", lambda *a, **kw: None)
+        monkeypatch.setattr(onboard_wizard, "console", SimpleNamespace(clear=lambda: None))
+
+        _configure_model_presets(config)
+
+        assert "old" not in config.model_presets
+        assert "old" not in _MODEL_PRESET_CACHE
+        _MODEL_PRESET_CACHE.clear()
+
+    def test_model_preset_field_handler(self, monkeypatch):
+        """_handle_model_preset_field should set a preset name from choices."""
+        from nanobot.cli.onboard import _MODEL_PRESET_CACHE, _handle_model_preset_field
+        from nanobot.config.schema import AgentDefaults
+
+        _MODEL_PRESET_CACHE.clear()
+        _MODEL_PRESET_CACHE.update({"fast", "power", "default"})
+
+        monkeypatch.setattr(onboard_wizard, "_select_with_back", lambda *a, **kw: "fast")
+
+        defaults = AgentDefaults()
+        _handle_model_preset_field(defaults, "model_preset", "Model Preset", None)
+        assert defaults.model_preset == "fast"
+        _MODEL_PRESET_CACHE.clear()
+
+    def test_model_preset_field_handler_clear(self, monkeypatch):
+        """_handle_model_preset_field should clear preset when (clear/unset) chosen."""
+        from nanobot.cli.onboard import _MODEL_PRESET_CACHE, _handle_model_preset_field
+        from nanobot.config.schema import AgentDefaults
+
+        _MODEL_PRESET_CACHE.clear()
+        _MODEL_PRESET_CACHE.add("fast")
+
+        monkeypatch.setattr(onboard_wizard, "_select_with_back", lambda *a, **kw: "(clear/unset)")
+
+        defaults = AgentDefaults(model_preset="fast")
+        _handle_model_preset_field(defaults, "model_preset", "Model Preset", "fast")
+        assert defaults.model_preset is None
+        _MODEL_PRESET_CACHE.clear()
+
+    def test_main_menu_dispatch_includes_model_presets(self):
+        """_configure_model_presets should be importable and callable."""
+        from nanobot.cli.onboard import _configure_model_presets
+
+        assert callable(_configure_model_presets)
+
+    def test_run_onboard_model_presets_edit(self, monkeypatch):
+        """run_onboard should handle [M] Model Presets correctly."""
+        from nanobot.config.schema import ModelPresetConfig
+
+        initial_config = Config()
+
+        responses = iter([
+            "[M] Model Presets",
+            "[S] Save and Exit",
+        ])
+
+        class FakePrompt:
+            def __init__(self, response):
+                self.response = response
+
+            def ask(self):
+                if isinstance(self.response, BaseException):
+                    raise self.response
+                return self.response
+
+        def fake_select(*_args, **_kwargs):
+            return FakePrompt(next(responses))
+
+        preset_mutated = {"n": 0}
+
+        def fake_configure_model_presets(config):
+            preset_mutated["n"] += 1
+            config.model_presets["test"] = ModelPresetConfig(model="gpt-test")
+
+        monkeypatch.setattr(onboard_wizard, "questionary", SimpleNamespace(select=fake_select))
+        monkeypatch.setattr(onboard_wizard, "_configure_model_presets", fake_configure_model_presets)
+        monkeypatch.setattr(onboard_wizard, "_show_main_menu_header", lambda: None)
+        monkeypatch.setattr(onboard_wizard, "_show_section_header", lambda *a, **kw: None)
+        monkeypatch.setattr(onboard_wizard, "console", SimpleNamespace(clear=lambda: None))
+
+        result = run_onboard(initial_config)
+        assert result.should_save is True
+        assert preset_mutated["n"] == 1
+        assert "test" in result.config.model_presets
+
+    def test_fallback_models_field_add(self, monkeypatch):
+        """_handle_fallback_models_field should add a preset name."""
+        from nanobot.cli.onboard import _MODEL_PRESET_CACHE, _handle_fallback_models_field
+        from nanobot.config.schema import AgentDefaults
+
+        _MODEL_PRESET_CACHE.clear()
+        _MODEL_PRESET_CACHE.update({"fast", "default"})
+
+        select_responses = iter(["fast"])
+        questionary_responses = iter(["[+] Add preset", "[Done]"])
+
+        class FakePrompt:
+            def __init__(self, response):
+                self.response = response
+
+            def ask(self):
+                if isinstance(self.response, BaseException):
+                    raise self.response
+                return self.response
+
+        def fake_questionary_select(*_args, **_kwargs):
+            return FakePrompt(next(questionary_responses))
+
+        def fake_select_with_back(*_args, **_kwargs):
+            return next(select_responses)
+
+        monkeypatch.setattr(
+            onboard_wizard, "questionary",
+            SimpleNamespace(select=fake_questionary_select, press_any_key_to_continue=lambda: FakePrompt(None)),
+        )
+        monkeypatch.setattr(onboard_wizard, "_select_with_back", fake_select_with_back)
+        monkeypatch.setattr(onboard_wizard, "console", SimpleNamespace(clear=lambda: None, print=lambda *a, **kw: None))
+
+        defaults = AgentDefaults()
+        _handle_fallback_models_field(defaults, "fallback_models", "Fallback Models", [])
+        assert defaults.fallback_models == ["fast"]
+        _MODEL_PRESET_CACHE.clear()
+
+    def test_provider_field_handler(self, monkeypatch):
+        """_handle_provider_field should set provider from choices."""
+        from nanobot.cli.onboard import _handle_provider_field
+        from nanobot.config.schema import AgentDefaults
+
+        monkeypatch.setattr(onboard_wizard, "_select_with_back", lambda *a, **kw: "anthropic")
+
+        defaults = AgentDefaults()
+        _handle_provider_field(defaults, "provider", "Provider", "auto")
+        assert defaults.provider == "anthropic"

tests/agent/test_runtime_refresh.py

@@ -47,3 +47,28 @@ def test_provider_refresh_updates_all_model_dependents(tmp_path: Path) -> None:
     assert loop.dream.provider is new_provider
     assert loop.dream.model == "new-model"
     assert loop.dream._runner.provider is new_provider
+
+
+def test_llm_runtime_refreshes_provider_snapshot(tmp_path: Path) -> None:
+    old_provider = _provider("old-model")
+    new_provider = _provider("new-model", max_tokens=456)
+    loop = AgentLoop(
+        bus=MessageBus(),
+        provider=old_provider,
+        workspace=tmp_path,
+        model="old-model",
+        context_window_tokens=1000,
+        provider_snapshot_loader=lambda: ProviderSnapshot(
+            provider=new_provider,
+            model="new-model",
+            context_window_tokens=2000,
+            signature=("new-model",),
+        ),
+    )
+
+    runtime = loop.llm_runtime()
+
+    assert runtime.provider is new_provider
+    assert runtime.model == "new-model"
+    assert loop.provider is new_provider
+    assert loop.runner.provider is new_provider

tests/agent/test_unified_session.py

@@ -387,6 +387,7 @@ class TestConsolidationUnaffectedByUnifiedSession:
 
         session = Session(key="unified:default")
         session.messages = [{"role": "user", "content": "msg"}]
+        sessions.get_or_create.return_value = session
 
         # Simulate over-budget: estimated > budget
         consolidator.estimate_session_prompt_tokens = MagicMock(return_value=(950, "tiktoken"))

tests/channels/test_websocket_channel.py

@@ -370,6 +370,55 @@ async def test_send_progress_includes_structured_tool_events() -> None:
     ]
 
 
+@pytest.mark.asyncio
+async def test_send_file_edit_progress_uses_file_edit_event() -> None:
+    bus = MagicMock()
+    channel = WebSocketChannel({"enabled": True, "allowFrom": ["*"]}, bus)
+    mock_ws = AsyncMock()
+    channel._attach(mock_ws, "chat-1")
+
+    await channel.send(OutboundMessage(
+        channel="websocket",
+        chat_id="chat-1",
+        content="",
+        metadata={
+            "_progress": True,
+            "_file_edit_events": [
+                {
+                    "version": 1,
+                    "phase": "start",
+                    "call_id": "call-1",
+                    "tool": "write_file",
+                    "path": "src/app.py",
+                    "added": 12,
+                    "deleted": 2,
+                    "approximate": True,
+                    "status": "editing",
+                }
+            ],
+        },
+    ))
+
+    payload = json.loads(mock_ws.send.await_args.args[0])
+    assert payload == {
+        "event": "file_edit",
+        "chat_id": "chat-1",
+        "edits": [
+            {
+                "version": 1,
+                "phase": "start",
+                "call_id": "call-1",
+                "tool": "write_file",
+                "path": "src/app.py",
+                "added": 12,
+                "deleted": 2,
+                "approximate": True,
+                "status": "editing",
+            }
+        ],
+    }
+
+
 @pytest.mark.asyncio
 async def test_send_progress_includes_agent_ui_blob() -> None:
     bus = MagicMock()
@@ -758,6 +807,25 @@ async def test_send_session_updated_emits_session_updated_event() -> None:
     assert body == {"event": "session_updated", "chat_id": "chat-1"}
 
 
+@pytest.mark.asyncio
+async def test_send_session_updated_includes_scope_when_present() -> None:
+    bus = MagicMock()
+    channel = WebSocketChannel({"enabled": True, "allowFrom": ["*"]}, bus)
+    mock_ws = AsyncMock()
+    channel._attach(mock_ws, "chat-1")
+
+    await channel.send(OutboundMessage(
+        channel="websocket",
+        chat_id="chat-1",
+        content="",
+        metadata={"_session_updated": True, "_session_update_scope": "metadata"},
+    ))
+
+    mock_ws.send.assert_awaited_once()
+    body = json.loads(mock_ws.send.await_args.args[0])
+    assert body == {"event": "session_updated", "chat_id": "chat-1", "scope": "metadata"}
+
+
 @pytest.mark.asyncio
 async def test_send_non_connection_closed_exception_is_raised() -> None:
     bus = MagicMock()
@@ -946,7 +1014,12 @@ async def test_settings_api_returns_safe_subset_and_updates_whitelist(
         providers = {provider["name"]: provider for provider in body["providers"]}
         assert providers["openai"]["configured"] is True
         assert providers["openai"]["api_key_hint"] == "secr••••-key"
+        assert providers["azure_openai"]["api_key_required"] is True
         assert providers["openrouter"]["configured"] is False
+        assert providers["openrouter"]["api_key_required"] is True
+        assert providers["atomic_chat"]["configured"] is False
+        assert providers["atomic_chat"]["api_key_required"] is False
+        assert providers["atomic_chat"]["default_api_base"] == "http://localhost:1337/v1"
         assert body["agent"]["has_api_key"] is True
         assert body["web_search"]["provider"] == "brave"
         assert body["web_search"]["api_key_hint"] == "brav••••cret"
@@ -969,10 +1042,24 @@ async def test_settings_api_returns_safe_subset_and_updates_whitelist(
         assert provider_rows["openrouter"]["configured"] is True
         assert "sk-or-test" not in provider_updated.text
 
+        local_provider_updated = await _http_get(
+            "http://127.0.0.1:"
+            f"{port}/api/settings/provider/update?provider=atomic_chat"
+            "&api_base=http%3A%2F%2Flocalhost%3A1337%2Fv1",
+            headers={"Authorization": "Bearer tok"},
+        )
+        assert local_provider_updated.status_code == 200
+        local_provider_body = local_provider_updated.json()
+        local_provider_rows = {
+            provider["name"]: provider for provider in local_provider_body["providers"]
+        }
+        assert local_provider_rows["atomic_chat"]["configured"] is True
+        assert "localhost:1337" in local_provider_updated.text
+
         updated = await _http_get(
             "http://127.0.0.1:"
-            f"{port}/api/settings/update?model=openrouter/test"
-            "&provider=openrouter",
+            f"{port}/api/settings/update?model=atomic_chat/test"
+            "&provider=atomic_chat",
             headers={"Authorization": "Bearer tok"},
         )
         assert updated.status_code == 200
@@ -992,10 +1079,11 @@ async def test_settings_api_returns_safe_subset_and_updates_whitelist(
         assert search_body["web_search"]["base_url"] == "https://search.example.com"
 
         saved = load_config(config_path)
-        assert saved.agents.defaults.model == "openrouter/test"
-        assert saved.agents.defaults.provider == "openrouter"
+        assert saved.agents.defaults.model == "atomic_chat/test"
+        assert saved.agents.defaults.provider == "atomic_chat"
         assert saved.providers.openrouter.api_key == "sk-or-test"
         assert saved.providers.openrouter.api_base == "https://openrouter.ai/api/v1"
+        assert saved.providers.atomic_chat.api_base == "http://localhost:1337/v1"
         assert saved.tools.web.search.provider == "searxng"
         assert saved.tools.web.search.api_key == ""
         assert saved.tools.web.search.base_url == "https://search.example.com"

tests/cli/test_commands.py

@@ -1170,6 +1170,7 @@ def test_gateway_cron_evaluator_receives_scheduled_reminder_context(
             self.model = "test-model"
             self.provider = kwargs.get("provider", object())
             self.tools = {}
+            seen["agent"] = self
 
         async def process_direct(self, *_args, **_kwargs):
             return OutboundMessage(
@@ -1218,6 +1219,11 @@ def test_gateway_cron_evaluator_receives_scheduled_reminder_context(
     assert isinstance(cron, _FakeCron)
     assert cron.on_job is not None
 
+    runtime_provider = object()
+    agent = seen["agent"]
+    agent.provider = runtime_provider
+    agent.model = "runtime-model"
+
     job = CronJob(
         id="cron-1",
         name="stretch",
@@ -1233,8 +1239,8 @@ def test_gateway_cron_evaluator_receives_scheduled_reminder_context(
 
     assert response == "Time to stretch."
     assert seen["response"] == "Time to stretch."
-    assert seen["provider"] is provider
-    assert seen["model"] == "test-model"
+    assert seen["provider"] is runtime_provider
+    assert seen["model"] == "runtime-model"
     assert seen["task_context"] == (
         "The scheduled time has arrived. Deliver this reminder to the user now, "
         "as a brief and natural message in their language. Speak directly to them — "
@@ -1543,6 +1549,9 @@ def test_gateway_health_endpoint_binds_and_serves_expected_responses(
             self.dream = _FakeDream()
             self.sessions = _FakeSessionManager()
 
+        def llm_runtime(self) -> None:
+            return None
+
         async def run(self) -> None:
             await asyncio.Event().wait()
 

tests/cli/test_interactive_retry_wait.py

@@ -69,6 +69,72 @@ async def test_reasoning_delta_displayed_when_show_reasoning_enabled():
     assert calls == ["I should search first."]
 
 
+@pytest.mark.asyncio
+async def test_reasoning_delta_buffers_until_sentence_boundary():
+    calls: list[str] = []
+    channels_config = SimpleNamespace(
+        send_progress=True, send_tool_hints=False, show_reasoning=True,
+    )
+    reasoning_buffer = commands._ReasoningBuffer()
+
+    with patch("nanobot.cli.commands._print_cli_reasoning", side_effect=lambda t, th, r=None: calls.append(t)):
+        first = await commands._maybe_print_interactive_progress(
+            SimpleNamespace(
+                content="The",
+                metadata={"_progress": True, "_reasoning_delta": True},
+            ),
+            None,
+            channels_config,
+            reasoning_buffer=reasoning_buffer,
+        )
+        second = await commands._maybe_print_interactive_progress(
+            SimpleNamespace(
+                content=" user asked.",
+                metadata={"_progress": True, "_reasoning_delta": True},
+            ),
+            None,
+            channels_config,
+            reasoning_buffer=reasoning_buffer,
+        )
+
+    assert first is True
+    assert second is True
+    assert calls == ["The user asked."]
+
+
+@pytest.mark.asyncio
+async def test_reasoning_end_flushes_buffered_delta():
+    calls: list[str] = []
+    channels_config = SimpleNamespace(
+        send_progress=True, send_tool_hints=False, show_reasoning=True,
+    )
+    reasoning_buffer = commands._ReasoningBuffer()
+
+    with patch("nanobot.cli.commands._print_cli_reasoning", side_effect=lambda t, th, r=None: calls.append(t)):
+        delta = await commands._maybe_print_interactive_progress(
+            SimpleNamespace(
+                content="The user asked",
+                metadata={"_progress": True, "_reasoning_delta": True},
+            ),
+            None,
+            channels_config,
+            reasoning_buffer=reasoning_buffer,
+        )
+        end = await commands._maybe_print_interactive_progress(
+            SimpleNamespace(
+                content="",
+                metadata={"_progress": True, "_reasoning_end": True},
+            ),
+            None,
+            channels_config,
+            reasoning_buffer=reasoning_buffer,
+        )
+
+    assert delta is True
+    assert end is True
+    assert calls == ["The user asked"]
+
+
 @pytest.mark.asyncio
 async def test_reasoning_hidden_when_show_reasoning_disabled():
     """Reasoning content should be suppressed when show_reasoning is False."""

tests/utils/test_file_edit_events.py

@@ -0,0 +1,83 @@
+from __future__ import annotations
+
+from pathlib import Path
+
+from nanobot.utils.file_edit_events import (
+    build_file_edit_end_event,
+    build_file_edit_start_event,
+    line_diff_stats,
+    prepare_file_edit_tracker,
+    read_file_snapshot,
+)
+
+
+def test_line_diff_stats_counts_replacements_insertions_and_deletions() -> None:
+    added, deleted = line_diff_stats("a\nb\nc\n", "a\nB\nc\nd\n")
+    assert (added, deleted) == (2, 1)
+
+
+def test_line_diff_stats_normalizes_crlf() -> None:
+    assert line_diff_stats("a\r\nb\r\n", "a\nb\nc\n") == (1, 0)
+
+
+def test_write_file_start_predicts_and_end_calibrates_exact_diff(tmp_path: Path) -> None:
+    target = tmp_path / "notes.txt"
+    target.write_text("old\nkeep\n", encoding="utf-8")
+    params = {"path": "notes.txt", "content": "new\nkeep\nextra\n"}
+    tracker = prepare_file_edit_tracker(
+        call_id="call-write",
+        tool_name="write_file",
+        tool=None,
+        workspace=tmp_path,
+        params=params,
+    )
+
+    assert tracker is not None
+    start = build_file_edit_start_event(tracker, params)
+    assert start == {
+        "version": 1,
+        "call_id": "call-write",
+        "tool": "write_file",
+        "path": "notes.txt",
+        "phase": "start",
+        "added": 2,
+        "deleted": 1,
+        "approximate": True,
+        "status": "editing",
+    }
+
+    target.write_text("new\nkeep\nextra\n", encoding="utf-8")
+    end = build_file_edit_end_event(tracker)
+    assert end["phase"] == "end"
+    assert end["status"] == "done"
+    assert end["approximate"] is False
+    assert (end["added"], end["deleted"]) == (2, 1)
+
+
+def test_binary_file_is_reported_but_not_counted(tmp_path: Path) -> None:
+    target = tmp_path / "data.bin"
+    target.write_bytes(b"\x00\x01before")
+    tracker = prepare_file_edit_tracker(
+        call_id="call-bin",
+        tool_name="edit_file",
+        tool=None,
+        workspace=tmp_path,
+        params={"path": "data.bin", "old_text": "before", "new_text": "after"},
+    )
+
+    assert tracker is not None
+    assert not read_file_snapshot(target).countable
+    target.write_bytes(b"\x00\x01after")
+    event = build_file_edit_end_event(tracker)
+    assert event["binary"] is True
+    assert (event["added"], event["deleted"]) == (0, 0)
+
+
+def test_untracked_tools_do_not_prepare_file_edit_tracker(tmp_path: Path) -> None:
+    assert prepare_file_edit_tracker(
+        call_id="call-exec",
+        tool_name="exec",
+        tool=None,
+        workspace=tmp_path,
+        params={"path": "created-by-shell.txt"},
+    ) is None

tests/utils/test_webui_transcript.py

@@ -42,6 +42,62 @@ def test_replay_delta_and_turn_end(tmp_path, monkeypatch) -> None:
     assert msgs[1]["latencyMs"] == 42
 
 
+def test_replay_file_edit_event_creates_file_activity(tmp_path, monkeypatch) -> None:
+    monkeypatch.setattr("nanobot.config.paths.get_data_dir", lambda: tmp_path)
+    key = "websocket:t-file"
+    for ev in (
+        {"event": "user", "chat_id": "t-file", "text": "edit"},
+        {
+            "event": "message",
+            "chat_id": "t-file",
+            "text": 'write_file({"path":"foo.txt"})',
+            "kind": "tool_hint",
+        },
+        {
+            "event": "file_edit",
+            "chat_id": "t-file",
+            "edits": [
+                {
+                    "version": 1,
+                    "call_id": "call-write",
+                    "tool": "write_file",
+                    "path": "foo.txt",
+                    "phase": "end",
+                    "added": 2,
+                    "deleted": 1,
+                    "approximate": False,
+                    "status": "done",
+                },
+            ],
+        },
+    ):
+        append_transcript_object(key, ev)
+
+    msgs = replay_transcript_to_ui_messages(read_transcript_lines(key))
+
+    assert len(msgs) == 3
+    assert msgs[1]["kind"] == "trace"
+    assert msgs[1]["traces"] == ['write_file({"path":"foo.txt"})']
+    assert "fileEdits" not in msgs[1]
+    assert msgs[2]["kind"] == "trace"
+    assert msgs[2]["traces"] == []
+    assert msgs[2]["fileEdits"] == [
+        {
+            "version": 1,
+            "call_id": "call-write",
+            "tool": "write_file",
+            "path": "foo.txt",
+            "phase": "end",
+            "added": 2,
+            "deleted": 1,
+            "approximate": False,
+            "status": "done",
+        },
+    ]
+    assert msgs[2]["activitySegmentId"]
+    assert msgs[2]["activitySegmentId"] != msgs[1]["activitySegmentId"]
+
+
 def test_build_response_schema(monkeypatch, tmp_path) -> None:
     from nanobot.utils.webui_transcript import build_webui_thread_response
 

webui/README.md

@@ -8,15 +8,11 @@ on the same port.
 For the project overview, install guide, and general docs map, see the root
 [`README.md`](../README.md).
 
-## Current status
+## Just want to use the WebUI?
 
-> [!NOTE]
-> The standalone WebUI development workflow currently requires a source
-> checkout.
->
-> WebUI changes in the GitHub repository may land before they are included in
-> the next packaged release, so source installs and published package versions
-> are not yet guaranteed to move in lockstep.
+If you installed nanobot via `pip install nanobot-ai`, the WebUI is **already bundled** in the wheel. Enable the WebSocket channel in `~/.nanobot/config.json` and run `nanobot gateway` — see the root [`README.md`](../README.md#-webui) for the 3-step setup. You do **not** need anything in this directory.
+
+This `webui/` tree is for people **hacking on the WebUI itself** (UI changes, new components, styling, etc.).
 
 ## Layout
 
@@ -25,7 +21,7 @@ webui/                 source tree (this directory)
 nanobot/web/dist/      build output served by the gateway
 ```
 
-## Develop from source
+## Develop the WebUI (Vite HMR)
 
 ### 1. Install nanobot from source
 
@@ -35,6 +31,8 @@ From the repository root:
 pip install -e .
 ```
 
+> Editable installs intentionally **skip** the WebUI bundle step — Vite HMR is faster than rebuilding `dist/` on every change.
+
 ### 2. Enable the WebSocket channel
 
 In `~/.nanobot/config.json`:
@@ -63,8 +61,7 @@ bun run dev
 
 Then open `http://127.0.0.1:5173`.
 
-By default, the dev server proxies `/api`, `/webui`, `/auth`, and WebSocket
-traffic to `http://127.0.0.1:8765`.
+By default the dev server proxies `/api`, `/webui`, `/auth`, and WebSocket traffic to `http://127.0.0.1:8765`.
 
 If your gateway listens on a non-default port, point the dev server at it:
 
@@ -74,7 +71,7 @@ NANOBOT_API_URL=http://127.0.0.1:9000 bun run dev
 
 ### Access from another device (LAN)
 
-To use the webui from another device on the same network, set `host` to `"0.0.0.0"` and configure a `token` or `tokenIssueSecret` in `~/.nanobot/config.json`:
+To use the WebUI from another device on the same network, set `host` to `"0.0.0.0"` and configure a `token` or `tokenIssueSecret` in `~/.nanobot/config.json`:
 
 ```json
 {
@@ -91,20 +88,20 @@ To use the webui from another device on the same network, set `host` to `"0.0.0.
 
 The gateway will refuse to start if `host` is `"0.0.0.0"` and neither `token` nor `tokenIssueSecret` is set.
 
-Then open `http://<your-ip>:8765` on the other device. The webui will show an authentication form where you enter the secret. It is saved in your browser so you only need to enter it once.
+Then open `http://<your-ip>:8765` on the other device. The WebUI will show an authentication form where you enter the secret. It is saved in your browser so you only need to enter it once.
 
 ## Build for packaged runtime
 
+You usually do not need to run this by hand: `python -m build` invokes the WebUI build automatically when packaging the wheel.
+
+If you want to preview the production bundle locally without rebuilding the wheel:
+
 ```bash
 cd webui
-bun run build
+bun run build          # writes to ../nanobot/web/dist
 ```
 
-This writes the production assets to `../nanobot/web/dist`, which is the
-directory served by `nanobot gateway` and bundled into the Python wheel.
-
-If you are cutting a release, run the build before packaging so the published
-wheel contains the current WebUI assets.
+The gateway picks up the new bundle on the next restart.
 
 ## Test
 

webui/bun.lock

@@ -15,12 +15,15 @@
         "@radix-ui/react-tooltip": "^1.1.6",
         "class-variance-authority": "^0.7.1",
         "clsx": "^2.1.1",
+        "i18next": "^26.0.6",
         "lucide-react": "^0.469.0",
         "react": "^18.3.1",
         "react-dom": "^18.3.1",
+        "react-i18next": "^17.0.4",
         "react-markdown": "^9.0.1",
         "react-syntax-highlighter": "^15.6.1",
         "rehype-katex": "^7.0.1",
+        "remark-breaks": "^4.0.0",
         "remark-gfm": "^4.0.0",
         "remark-math": "^6.0.0",
         "tailwind-merge": "^2.6.0",
@@ -506,8 +509,12 @@
 
     "highlightjs-vue": ["highlightjs-vue@1.0.0", "", {}, "sha512-PDEfEF102G23vHmPhLyPboFCD+BkMGu+GuJe2d9/eH4FsCwvgBpnc9n0pGE+ffKdph38s6foEZiEjdgHdzp+IA=="],
 
+    "html-parse-stringify": ["html-parse-stringify@3.0.1", "", { "dependencies": { "void-elements": "3.1.0" } }, "sha512-KknJ50kTInJ7qIScF3jeaFRpMpE8/lfiTdzf/twXyPBLAGrLRTmkz3AdTnKeh40X8k9L2fdYwEp/42WGXIRGcg=="],
+
     "html-url-attributes": ["html-url-attributes@3.0.1", "", {}, "sha512-ol6UPyBWqsrO6EJySPz2O7ZSr856WDrEzM5zMqp+FJJLGMW35cLYmmZnl0vztAZxRUoNZJFTCohfjuIJ8I4QBQ=="],
 
+    "i18next": ["i18next@26.2.0", "", { "peerDependencies": { "typescript": "^5 || ^6" }, "optionalPeers": ["typescript"] }, "sha512-zwBHldHdTmwN7r6UNc7lC6GWNN+YYg3DrRSeHR5PRRBf5QnJZcYHrQc0uaU26qZeYxR7iFZD+Y315dPnKP47wA=="],
+
     "indent-string": ["indent-string@4.0.0", "", {}, "sha512-EdDDZu4A2OyIK7Lr/2zG+w5jmbuk1DVBnEwREQvBzspBJkCEbRa8GxU1lghYcaGJCnRWibjDXlq779X1/y5xwg=="],
 
     "inline-style-parser": ["inline-style-parser@0.2.7", "", {}, "sha512-Nb2ctOyNR8DqQoR0OwRG95uNWIC0C1lCgf5Naz5H6Ji72KZ8OcFZLz2P5sNgwlyoJ8Yif11oMuYs5pBQa86csA=="],
@@ -588,6 +595,8 @@
 
     "mdast-util-mdxjs-esm": ["mdast-util-mdxjs-esm@2.0.1", "", { "dependencies": { "@types/estree-jsx": "^1.0.0", "@types/hast": "^3.0.0", "@types/mdast": "^4.0.0", "devlop": "^1.0.0", "mdast-util-from-markdown": "^2.0.0", "mdast-util-to-markdown": "^2.0.0" } }, "sha512-EcmOpxsZ96CvlP03NghtH1EsLtr0n9Tm4lPUJUBccV9RwUOneqSycg19n5HGzCf+10LozMRSObtVr3ee1WoHtg=="],
 
+    "mdast-util-newline-to-break": ["mdast-util-newline-to-break@2.0.0", "", { "dependencies": { "@types/mdast": "^4.0.0", "mdast-util-find-and-replace": "^3.0.0" } }, "sha512-MbgeFca0hLYIEx/2zGsszCSEJJ1JSCdiY5xQxRcLDDGa8EPvlLPupJ4DSajbMPAnC0je8jfb9TiUATnxxrHUog=="],
+
     "mdast-util-phrasing": ["mdast-util-phrasing@4.1.0", "", { "dependencies": { "@types/mdast": "^4.0.0", "unist-util-is": "^6.0.0" } }, "sha512-TqICwyvJJpBwvGAMZjj4J2n0X8QWp21b9l0o7eXyVJ25YNWYbJDVIyD1bZXE6WtV6RmKJVYmQAKWa0zWOABz2w=="],
 
     "mdast-util-to-hast": ["mdast-util-to-hast@13.2.1", "", { "dependencies": { "@types/hast": "^3.0.0", "@types/mdast": "^4.0.0", "@ungap/structured-clone": "^1.0.0", "devlop": "^1.0.0", "micromark-util-sanitize-uri": "^2.0.0", "trim-lines": "^3.0.0", "unist-util-position": "^5.0.0", "unist-util-visit": "^5.0.0", "vfile": "^6.0.0" } }, "sha512-cctsq2wp5vTsLIcaymblUriiTcZd0CwWtCbLvrOzYCDZoWyMNV8sZ7krj09FSnsiJi3WVsHLM4k6Dq/yaPyCXA=="],
@@ -718,6 +727,8 @@
 
     "react-dom": ["react-dom@18.3.1", "", { "dependencies": { "loose-envify": "^1.1.0", "scheduler": "^0.23.2" }, "peerDependencies": { "react": "^18.3.1" } }, "sha512-5m4nQKp+rZRb09LNH59GM4BxTh9251/ylbKIbpe7TpGxfJ+9kv6BLkLBXIjjspbgbnIBNqlI23tRnTWT0snUIw=="],
 
+    "react-i18next": ["react-i18next@17.0.8", "", { "dependencies": { "@babel/runtime": "^7.29.2", "html-parse-stringify": "^3.0.1", "use-sync-external-store": "^1.6.0" }, "peerDependencies": { "i18next": ">= 26.2.0", "react": ">= 16.8.0", "typescript": "^5 || ^6" }, "optionalPeers": ["typescript"] }, "sha512-0ooKbGLU8JXhe1zwpQUWIeXSgLPOfwJmgheWRIUpcoA0CpyabpGhayjdG+/eA5esC1AQ8h2jWpXjJfzQzeDOCw=="],
+
     "react-is": ["react-is@17.0.2", "", {}, "sha512-w2GsyukL62IJnlaff/nRegPQR94C/XXamvMWmSHRJ4y7Ts/4ocGRmTHvOs8PSE6pB3dWOrD/nueuU5sduBsQ4w=="],
 
     "react-markdown": ["react-markdown@9.1.0", "", { "dependencies": { "@types/hast": "^3.0.0", "@types/mdast": "^4.0.0", "devlop": "^1.0.0", "hast-util-to-jsx-runtime": "^2.0.0", "html-url-attributes": "^3.0.0", "mdast-util-to-hast": "^13.0.0", "remark-parse": "^11.0.0", "remark-rehype": "^11.0.0", "unified": "^11.0.0", "unist-util-visit": "^5.0.0", "vfile": "^6.0.0" }, "peerDependencies": { "@types/react": ">=18", "react": ">=18" } }, "sha512-xaijuJB0kzGiUdG7nc2MOMDUDBWPyGAjZtUrow9XxUeua8IqeP+VlIfAZ3bphpcLTnSZXz6z9jcVC/TCwbfgdw=="],
@@ -742,6 +753,8 @@
 
     "rehype-katex": ["rehype-katex@7.0.1", "", { "dependencies": { "@types/hast": "^3.0.0", "@types/katex": "^0.16.0", "hast-util-from-html-isomorphic": "^2.0.0", "hast-util-to-text": "^4.0.0", "katex": "^0.16.0", "unist-util-visit-parents": "^6.0.0", "vfile": "^6.0.0" } }, "sha512-OiM2wrZ/wuhKkigASodFoo8wimG3H12LWQaH8qSPVJn9apWKFSH3YOCtbKpBorTVw/eI7cuT21XBbvwEswbIOA=="],
 
+    "remark-breaks": ["remark-breaks@4.0.0", "", { "dependencies": { "@types/mdast": "^4.0.0", "mdast-util-newline-to-break": "^2.0.0", "unified": "^11.0.0" } }, "sha512-IjEjJOkH4FuJvHZVIW0QCDWxcG96kCq7An/KVH2NfJe6rKZU2AsHeB3OEjPNRxi4QC34Xdx7I2KGYn6IpT7gxQ=="],
+
     "remark-gfm": ["remark-gfm@4.0.1", "", { "dependencies": { "@types/mdast": "^4.0.0", "mdast-util-gfm": "^3.0.0", "micromark-extension-gfm": "^3.0.0", "remark-parse": "^11.0.0", "remark-stringify": "^11.0.0", "unified": "^11.0.0" } }, "sha512-1quofZ2RQ9EWdeN34S79+KExV1764+wCUGop5CPL1WGdD0ocPpu91lzPGbwWMECpEpd42kJGQwzRfyov9j4yNg=="],
 
     "remark-math": ["remark-math@6.0.0", "", { "dependencies": { "@types/mdast": "^4.0.0", "mdast-util-math": "^3.0.0", "micromark-extension-math": "^3.0.0", "unified": "^11.0.0" } }, "sha512-MMqgnP74Igy+S3WwnhQ7kqGlEerTETXMvJhrUzDikVZ2/uogJCb+WHUg97hK9/jcfc0dkD73s3LN8zU49cTEtA=="],
@@ -860,6 +873,8 @@
 
     "vitest": ["vitest@2.1.9", "", { "dependencies": { "@vitest/expect": "2.1.9", "@vitest/mocker": "2.1.9", "@vitest/pretty-format": "^2.1.9", "@vitest/runner": "2.1.9", "@vitest/snapshot": "2.1.9", "@vitest/spy": "2.1.9", "@vitest/utils": "2.1.9", "chai": "^5.1.2", "debug": "^4.3.7", "expect-type": "^1.1.0", "magic-string": "^0.30.12", "pathe": "^1.1.2", "std-env": "^3.8.0", "tinybench": "^2.9.0", "tinyexec": "^0.3.1", "tinypool": "^1.0.1", "tinyrainbow": "^1.2.0", "vite": "^5.0.0", "vite-node": "2.1.9", "why-is-node-running": "^2.3.0" }, "peerDependencies": { "@edge-runtime/vm": "*", "@types/node": "^18.0.0 || >=20.0.0", "@vitest/browser": "2.1.9", "@vitest/ui": "2.1.9", "happy-dom": "*", "jsdom": "*" }, "optionalPeers": ["@edge-runtime/vm", "@types/node", "@vitest/browser", "@vitest/ui", "happy-dom", "jsdom"], "bin": { "vitest": "vitest.mjs" } }, "sha512-MSmPM9REYqDGBI8439mA4mWhV5sKmDlBKWIYbA3lRb2PTHACE0mgKwA8yQ2xq9vxDTuk4iPrECBAEW2aoFXY0Q=="],
 
+    "void-elements": ["void-elements@3.1.0", "", {}, "sha512-Dhxzh5HZuiHQhbvTW9AMetFfBHDMYpo23Uo9btPXgdYP+3T5S+p+jgNy7spra+veYhBP2dCSgxR/i2Y02h5/6w=="],
+
     "web-namespaces": ["web-namespaces@2.0.1", "", {}, "sha512-bKr1DkiNa2krS7qxNtdrtHAmzuYGFQLiQ13TsorsdT6ULTkPLKuu5+GsFpDlg6JFjUTwX2DyhMPG2be8uPrqsQ=="],
 
     "webidl-conversions": ["webidl-conversions@7.0.0", "", {}, "sha512-VwddBukDzu71offAQR975unBIGqfKZpM+8ZX6ySk8nYhVoo5CYaZyzt3YBvYtRtO+aoGlqxPg/B87NGVZ/fu6g=="],

webui/package-lock.json

@@ -318,6 +318,278 @@
         "node": ">=6.9.0"
       }
     },
+    "node_modules/@esbuild/aix-ppc64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/aix-ppc64/-/aix-ppc64-0.21.5.tgz",
+      "integrity": "sha512-1SDgH6ZSPTlggy1yI6+Dbkiz8xzpHJEVAlF/AM1tHPLsf5STom9rwtjE4hKAF20FfXXNTFqEYXyJNWh1GiZedQ==",
+      "cpu": [
+        "ppc64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "aix"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/android-arm": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/android-arm/-/android-arm-0.21.5.tgz",
+      "integrity": "sha512-vCPvzSjpPHEi1siZdlvAlsPxXl7WbOVUBBAowWug4rJHb68Ox8KualB+1ocNvT5fjv6wpkX6o/iEpbDrf68zcg==",
+      "cpu": [
+        "arm"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "android"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/android-arm64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/android-arm64/-/android-arm64-0.21.5.tgz",
+      "integrity": "sha512-c0uX9VAUBQ7dTDCjq+wdyGLowMdtR/GoC2U5IYk/7D1H1JYC0qseD7+11iMP2mRLN9RcCMRcjC4YMclCzGwS/A==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "android"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/android-x64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/android-x64/-/android-x64-0.21.5.tgz",
+      "integrity": "sha512-D7aPRUUNHRBwHxzxRvp856rjUHRFW1SdQATKXH2hqA0kAZb1hKmi02OpYRacl0TxIGz/ZmXWlbZgjwWYaCakTA==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "android"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/darwin-arm64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/darwin-arm64/-/darwin-arm64-0.21.5.tgz",
+      "integrity": "sha512-DwqXqZyuk5AiWWf3UfLiRDJ5EDd49zg6O9wclZ7kUMv2WRFr4HKjXp/5t8JZ11QbQfUS6/cRCKGwYhtNAY88kQ==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "darwin"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/darwin-x64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/darwin-x64/-/darwin-x64-0.21.5.tgz",
+      "integrity": "sha512-se/JjF8NlmKVG4kNIuyWMV/22ZaerB+qaSi5MdrXtd6R08kvs2qCN4C09miupktDitvh8jRFflwGFBQcxZRjbw==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "darwin"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/freebsd-arm64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/freebsd-arm64/-/freebsd-arm64-0.21.5.tgz",
+      "integrity": "sha512-5JcRxxRDUJLX8JXp/wcBCy3pENnCgBR9bN6JsY4OmhfUtIHe3ZW0mawA7+RDAcMLrMIZaf03NlQiX9DGyB8h4g==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "freebsd"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/freebsd-x64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/freebsd-x64/-/freebsd-x64-0.21.5.tgz",
+      "integrity": "sha512-J95kNBj1zkbMXtHVH29bBriQygMXqoVQOQYA+ISs0/2l3T9/kj42ow2mpqerRBxDJnmkUDCaQT/dfNXWX/ZZCQ==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "freebsd"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/linux-arm": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-arm/-/linux-arm-0.21.5.tgz",
+      "integrity": "sha512-bPb5AHZtbeNGjCKVZ9UGqGwo8EUu4cLq68E95A53KlxAPRmUyYv2D6F0uUI65XisGOL1hBP5mTronbgo+0bFcA==",
+      "cpu": [
+        "arm"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/linux-arm64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-arm64/-/linux-arm64-0.21.5.tgz",
+      "integrity": "sha512-ibKvmyYzKsBeX8d8I7MH/TMfWDXBF3db4qM6sy+7re0YXya+K1cem3on9XgdT2EQGMu4hQyZhan7TeQ8XkGp4Q==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/linux-ia32": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-ia32/-/linux-ia32-0.21.5.tgz",
+      "integrity": "sha512-YvjXDqLRqPDl2dvRODYmmhz4rPeVKYvppfGYKSNGdyZkA01046pLWyRKKI3ax8fbJoK5QbxblURkwK/MWY18Tg==",
+      "cpu": [
+        "ia32"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/linux-loong64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-loong64/-/linux-loong64-0.21.5.tgz",
+      "integrity": "sha512-uHf1BmMG8qEvzdrzAqg2SIG/02+4/DHB6a9Kbya0XDvwDEKCoC8ZRWI5JJvNdUjtciBGFQ5PuBlpEOXQj+JQSg==",
+      "cpu": [
+        "loong64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/linux-mips64el": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-mips64el/-/linux-mips64el-0.21.5.tgz",
+      "integrity": "sha512-IajOmO+KJK23bj52dFSNCMsz1QP1DqM6cwLUv3W1QwyxkyIWecfafnI555fvSGqEKwjMXVLokcV5ygHW5b3Jbg==",
+      "cpu": [
+        "mips64el"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/linux-ppc64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-ppc64/-/linux-ppc64-0.21.5.tgz",
+      "integrity": "sha512-1hHV/Z4OEfMwpLO8rp7CvlhBDnjsC3CttJXIhBi+5Aj5r+MBvy4egg7wCbe//hSsT+RvDAG7s81tAvpL2XAE4w==",
+      "cpu": [
+        "ppc64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/linux-riscv64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-riscv64/-/linux-riscv64-0.21.5.tgz",
+      "integrity": "sha512-2HdXDMd9GMgTGrPWnJzP2ALSokE/0O5HhTUvWIbD3YdjME8JwvSCnNGBnTThKGEB91OZhzrJ4qIIxk/SBmyDDA==",
+      "cpu": [
+        "riscv64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/linux-s390x": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-s390x/-/linux-s390x-0.21.5.tgz",
+      "integrity": "sha512-zus5sxzqBJD3eXxwvjN1yQkRepANgxE9lgOW2qLnmr8ikMTphkjgXu1HR01K4FJg8h1kEEDAqDcZQtbrRnB41A==",
+      "cpu": [
+        "s390x"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
     "node_modules/@esbuild/linux-x64": {
       "version": "0.21.5",
       "cpu": [
@@ -333,6 +605,108 @@
         "node": ">=12"
       }
     },
+    "node_modules/@esbuild/netbsd-x64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/netbsd-x64/-/netbsd-x64-0.21.5.tgz",
+      "integrity": "sha512-Woi2MXzXjMULccIwMnLciyZH4nCIMpWQAs049KEeMvOcNADVxo0UBIQPfSmxB3CWKedngg7sWZdLvLczpe0tLg==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "netbsd"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/openbsd-x64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/openbsd-x64/-/openbsd-x64-0.21.5.tgz",
+      "integrity": "sha512-HLNNw99xsvx12lFBUwoT8EVCsSvRNDVxNpjZ7bPn947b8gJPzeHWyNVhFsaerc0n3TsbOINvRP2byTZ5LKezow==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "openbsd"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/sunos-x64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/sunos-x64/-/sunos-x64-0.21.5.tgz",
+      "integrity": "sha512-6+gjmFpfy0BHU5Tpptkuh8+uw3mnrvgs+dSPQXQOv3ekbordwnzTVEb4qnIvQcYXq6gzkyTnoZ9dZG+D4garKg==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "sunos"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/win32-arm64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/win32-arm64/-/win32-arm64-0.21.5.tgz",
+      "integrity": "sha512-Z0gOTd75VvXqyq7nsl93zwahcTROgqvuAcYDUr+vOv8uHhNSKROyU961kgtCD1e95IqPKSQKH7tBTslnS3tA8A==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "win32"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/win32-ia32": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/win32-ia32/-/win32-ia32-0.21.5.tgz",
+      "integrity": "sha512-SWXFF1CL2RVNMaVs+BBClwtfZSvDgtL//G/smwAc5oVK/UPu2Gu9tIaRgFmYFFKrmg3SyAjSrElf0TiJ1v8fYA==",
+      "cpu": [
+        "ia32"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "win32"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/win32-x64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/win32-x64/-/win32-x64-0.21.5.tgz",
+      "integrity": "sha512-tQd/1efJuzPC6rCFwEvLtci/xNFcTZknmXs98FYDfGE4wP9ClFV98nyKrzJKVPMhdDnjzLhdUyMX4PsQAPjwIw==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "win32"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
     "node_modules/@floating-ui/core": {
       "version": "1.7.5",
       "license": "MIT",
@@ -1280,6 +1654,277 @@
       "dev": true,
       "license": "MIT"
     },
+    "node_modules/@rollup/rollup-android-arm-eabi": {
+      "version": "4.60.1",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-android-arm-eabi/-/rollup-android-arm-eabi-4.60.1.tgz",
+      "integrity": "sha512-d6FinEBLdIiK+1uACUttJKfgZREXrF0Qc2SmLII7W2AD8FfiZ9Wjd+rD/iRuf5s5dWrr1GgwXCvPqOuDquOowA==",
+      "cpu": [
+        "arm"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "android"
+      ]
+    },
+    "node_modules/@rollup/rollup-android-arm64": {
+      "version": "4.60.1",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-android-arm64/-/rollup-android-arm64-4.60.1.tgz",
+      "integrity": "sha512-YjG/EwIDvvYI1YvYbHvDz/BYHtkY4ygUIXHnTdLhG+hKIQFBiosfWiACWortsKPKU/+dUwQQCKQM3qrDe8c9BA==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "android"
+      ]
+    },
+    "node_modules/@rollup/rollup-darwin-arm64": {
+      "version": "4.60.1",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-darwin-arm64/-/rollup-darwin-arm64-4.60.1.tgz",
+      "integrity": "sha512-mjCpF7GmkRtSJwon+Rq1N8+pI+8l7w5g9Z3vWj4T7abguC4Czwi3Yu/pFaLvA3TTeMVjnu3ctigusqWUfjZzvw==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "darwin"
+      ]
+    },
+    "node_modules/@rollup/rollup-darwin-x64": {
+      "version": "4.60.1",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-darwin-x64/-/rollup-darwin-x64-4.60.1.tgz",
+      "integrity": "sha512-haZ7hJ1JT4e9hqkoT9R/19XW2QKqjfJVv+i5AGg57S+nLk9lQnJ1F/eZloRO3o9Scy9CM3wQ9l+dkXtcBgN5Ew==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "darwin"
+      ]
+    },
+    "node_modules/@rollup/rollup-freebsd-arm64": {
+      "version": "4.60.1",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-freebsd-arm64/-/rollup-freebsd-arm64-4.60.1.tgz",
+      "integrity": "sha512-czw90wpQq3ZsAVBlinZjAYTKduOjTywlG7fEeWKUA7oCmpA8xdTkxZZlwNJKWqILlq0wehoZcJYfBvOyhPTQ6w==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "freebsd"
+      ]
+    },
+    "node_modules/@rollup/rollup-freebsd-x64": {
+      "version": "4.60.1",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-freebsd-x64/-/rollup-freebsd-x64-4.60.1.tgz",
+      "integrity": "sha512-KVB2rqsxTHuBtfOeySEyzEOB7ltlB/ux38iu2rBQzkjbwRVlkhAGIEDiiYnO2kFOkJp+Z7pUXKyrRRFuFUKt+g==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "freebsd"
+      ]
+    },
+    "node_modules/@rollup/rollup-linux-arm-gnueabihf": {
+      "version": "4.60.1",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-arm-gnueabihf/-/rollup-linux-arm-gnueabihf-4.60.1.tgz",
+      "integrity": "sha512-L+34Qqil+v5uC0zEubW7uByo78WOCIrBvci69E7sFASRl0X7b/MB6Cqd1lky/CtcSVTydWa2WZwFuWexjS5o6g==",
+      "cpu": [
+        "arm"
+      ],
+      "dev": true,
+      "libc": [
+        "glibc"
+      ],
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
+    "node_modules/@rollup/rollup-linux-arm-musleabihf": {
+      "version": "4.60.1",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-arm-musleabihf/-/rollup-linux-arm-musleabihf-4.60.1.tgz",
+      "integrity": "sha512-n83O8rt4v34hgFzlkb1ycniJh7IR5RCIqt6mz1VRJD6pmhRi0CXdmfnLu9dIUS6buzh60IvACM842Ffb3xd6Gg==",
+      "cpu": [
+        "arm"
+      ],
+      "dev": true,
+      "libc": [
+        "musl"
+      ],
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
+    "node_modules/@rollup/rollup-linux-arm64-gnu": {
+      "version": "4.60.1",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-arm64-gnu/-/rollup-linux-arm64-gnu-4.60.1.tgz",
+      "integrity": "sha512-Nql7sTeAzhTAja3QXeAI48+/+GjBJ+QmAH13snn0AJSNL50JsDqotyudHyMbO2RbJkskbMbFJfIJKWA6R1LCJQ==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "libc": [
+        "glibc"
+      ],
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
+    "node_modules/@rollup/rollup-linux-arm64-musl": {
+      "version": "4.60.1",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-arm64-musl/-/rollup-linux-arm64-musl-4.60.1.tgz",
+      "integrity": "sha512-+pUymDhd0ys9GcKZPPWlFiZ67sTWV5UU6zOJat02M1+PiuSGDziyRuI/pPue3hoUwm2uGfxdL+trT6Z9rxnlMA==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "libc": [
+        "musl"
+      ],
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
+    "node_modules/@rollup/rollup-linux-loong64-gnu": {
+      "version": "4.60.1",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-loong64-gnu/-/rollup-linux-loong64-gnu-4.60.1.tgz",
+      "integrity": "sha512-VSvgvQeIcsEvY4bKDHEDWcpW4Yw7BtlKG1GUT4FzBUlEKQK0rWHYBqQt6Fm2taXS+1bXvJT6kICu5ZwqKCnvlQ==",
+      "cpu": [
+        "loong64"
+      ],
+      "dev": true,
+      "libc": [
+        "glibc"
+      ],
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
+    "node_modules/@rollup/rollup-linux-loong64-musl": {
+      "version": "4.60.1",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-loong64-musl/-/rollup-linux-loong64-musl-4.60.1.tgz",
+      "integrity": "sha512-4LqhUomJqwe641gsPp6xLfhqWMbQV04KtPp7/dIp0nzPxAkNY1AbwL5W0MQpcalLYk07vaW9Kp1PBhdpZYYcEw==",
+      "cpu": [
+        "loong64"
+      ],
+      "dev": true,
+      "libc": [
+        "musl"
+      ],
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
+    "node_modules/@rollup/rollup-linux-ppc64-gnu": {
+      "version": "4.60.1",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-ppc64-gnu/-/rollup-linux-ppc64-gnu-4.60.1.tgz",
+      "integrity": "sha512-tLQQ9aPvkBxOc/EUT6j3pyeMD6Hb8QF2BTBnCQWP/uu1lhc9AIrIjKnLYMEroIz/JvtGYgI9dF3AxHZNaEH0rw==",
+      "cpu": [
+        "ppc64"
+      ],
+      "dev": true,
+      "libc": [
+        "glibc"
+      ],
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
+    "node_modules/@rollup/rollup-linux-ppc64-musl": {
+      "version": "4.60.1",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-ppc64-musl/-/rollup-linux-ppc64-musl-4.60.1.tgz",
+      "integrity": "sha512-RMxFhJwc9fSXP6PqmAz4cbv3kAyvD1etJFjTx4ONqFP9DkTkXsAMU4v3Vyc5BgzC+anz7nS/9tp4obsKfqkDHg==",
+      "cpu": [
+        "ppc64"
+      ],
+      "dev": true,
+      "libc": [
+        "musl"
+      ],
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
+    "node_modules/@rollup/rollup-linux-riscv64-gnu": {
+      "version": "4.60.1",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-riscv64-gnu/-/rollup-linux-riscv64-gnu-4.60.1.tgz",
+      "integrity": "sha512-QKgFl+Yc1eEk6MmOBfRHYF6lTxiiiV3/z/BRrbSiW2I7AFTXoBFvdMEyglohPj//2mZS4hDOqeB0H1ACh3sBbg==",
+      "cpu": [
+        "riscv64"
+      ],
+      "dev": true,
+      "libc": [
+        "glibc"
+      ],
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
+    "node_modules/@rollup/rollup-linux-riscv64-musl": {
+      "version": "4.60.1",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-riscv64-musl/-/rollup-linux-riscv64-musl-4.60.1.tgz",
+      "integrity": "sha512-RAjXjP/8c6ZtzatZcA1RaQr6O1TRhzC+adn8YZDnChliZHviqIjmvFwHcxi4JKPSDAt6Uhf/7vqcBzQJy0PDJg==",
+      "cpu": [
+        "riscv64"
+      ],
+      "dev": true,
+      "libc": [
+        "musl"
+      ],
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
+    "node_modules/@rollup/rollup-linux-s390x-gnu": {
+      "version": "4.60.1",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-s390x-gnu/-/rollup-linux-s390x-gnu-4.60.1.tgz",
+      "integrity": "sha512-wcuocpaOlaL1COBYiA89O6yfjlp3RwKDeTIA0hM7OpmhR1Bjo9j31G1uQVpDlTvwxGn2nQs65fBFL5UFd76FcQ==",
+      "cpu": [
+        "s390x"
+      ],
+      "dev": true,
+      "libc": [
+        "glibc"
+      ],
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
     "node_modules/@rollup/rollup-linux-x64-gnu": {
       "version": "4.60.1",
       "cpu": [
@@ -1304,6 +1949,90 @@
         "linux"
       ]
     },
+    "node_modules/@rollup/rollup-openbsd-x64": {
+      "version": "4.60.1",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-openbsd-x64/-/rollup-openbsd-x64-4.60.1.tgz",
+      "integrity": "sha512-cl0w09WsCi17mcmWqqglez9Gk8isgeWvoUZ3WiJFYSR3zjBQc2J5/ihSjpl+VLjPqjQ/1hJRcqBfLjssREQILw==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "openbsd"
+      ]
+    },
+    "node_modules/@rollup/rollup-openharmony-arm64": {
+      "version": "4.60.1",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-openharmony-arm64/-/rollup-openharmony-arm64-4.60.1.tgz",
+      "integrity": "sha512-4Cv23ZrONRbNtbZa37mLSueXUCtN7MXccChtKpUnQNgF010rjrjfHx3QxkS2PI7LqGT5xXyYs1a7LbzAwT0iCA==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "openharmony"
+      ]
+    },
+    "node_modules/@rollup/rollup-win32-arm64-msvc": {
+      "version": "4.60.1",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-win32-arm64-msvc/-/rollup-win32-arm64-msvc-4.60.1.tgz",
+      "integrity": "sha512-i1okWYkA4FJICtr7KpYzFpRTHgy5jdDbZiWfvny21iIKky5YExiDXP+zbXzm3dUcFpkEeYNHgQ5fuG236JPq0g==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "win32"
+      ]
+    },
+    "node_modules/@rollup/rollup-win32-ia32-msvc": {
+      "version": "4.60.1",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-win32-ia32-msvc/-/rollup-win32-ia32-msvc-4.60.1.tgz",
+      "integrity": "sha512-u09m3CuwLzShA0EYKMNiFgcjjzwqtUMLmuCJLeZWjjOYA3IT2Di09KaxGBTP9xVztWyIWjVdsB2E9goMjZvTQg==",
+      "cpu": [
+        "ia32"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "win32"
+      ]
+    },
+    "node_modules/@rollup/rollup-win32-x64-gnu": {
+      "version": "4.60.1",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-win32-x64-gnu/-/rollup-win32-x64-gnu-4.60.1.tgz",
+      "integrity": "sha512-k+600V9Zl1CM7eZxJgMyTUzmrmhB/0XZnF4pRypKAlAgxmedUA+1v9R+XOFv56W4SlHEzfeMtzujLJD22Uz5zg==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "win32"
+      ]
+    },
+    "node_modules/@rollup/rollup-win32-x64-msvc": {
+      "version": "4.60.1",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-win32-x64-msvc/-/rollup-win32-x64-msvc-4.60.1.tgz",
+      "integrity": "sha512-lWMnixq/QzxyhTV6NjQJ4SFo1J6PvOX8vUx5Wb4bBPsEb+8xZ89Bz6kOXpfXj9ak9AHTQVQzlgzBEc1SyM27xQ==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "win32"
+      ]
+    },
     "node_modules/@tailwindcss/typography": {
       "version": "0.5.19",
       "dev": true,
@@ -2309,6 +3038,21 @@
         "url": "https://github.com/sponsors/rawify"
       }
     },
+    "node_modules/fsevents": {
+      "version": "2.3.3",
+      "resolved": "https://registry.npmjs.org/fsevents/-/fsevents-2.3.3.tgz",
+      "integrity": "sha512-5xoDfX+fL7faATnagmWPpbFtwh/R77WmMMqqHGS65C3vvB0YHrgF+B1YmZ3441tMj5n63k0212XNoJwzlhffQw==",
+      "dev": true,
+      "hasInstallScript": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "darwin"
+      ],
+      "engines": {
+        "node": "^8.16.0 || ^10.6.0 || >=11.0.0"
+      }
+    },
     "node_modules/function-bind": {
       "version": "1.1.2",
       "dev": true,

webui/package.json

@@ -30,6 +30,7 @@
     "react-markdown": "^9.0.1",
     "react-syntax-highlighter": "^15.6.1",
     "rehype-katex": "^7.0.1",
+    "remark-breaks": "^4.0.0",
     "remark-gfm": "^4.0.0",
     "remark-math": "^6.0.0",
     "tailwind-merge": "^2.6.0"

webui/src/App.tsx

@@ -7,7 +7,8 @@ import { ThreadShell } from "@/components/thread/ThreadShell";
 import { Sheet, SheetContent } from "@/components/ui/sheet";
 
 import { useSessions } from "@/hooks/useSessions";
-import { useTheme } from "@/hooks/useTheme";
+import { useDeferredTitleRefresh } from "@/hooks/useDeferredTitleRefresh";
+import { ThemeProvider, useTheme } from "@/hooks/useTheme";
 import { cn } from "@/lib/utils";
 import {
   clearSavedSecret,
@@ -16,6 +17,7 @@ import {
   loadSavedSecret,
   saveSecret,
 } from "@/lib/bootstrap";
+import { deriveTitle } from "@/lib/format";
 import { NanobotClient } from "@/lib/nanobot-client";
 import { ClientProvider, useClient } from "@/providers/ClientProvider";
 import type { ChatSummary } from "@/lib/types";
@@ -30,14 +32,30 @@ type BootState =
       status: "ready";
       client: NanobotClient;
       token: string;
+      tokenExpiresAt: number;
       modelName: string | null;
     };
 
 const SIDEBAR_STORAGE_KEY = "nanobot-webui.sidebar";
 const RESTART_STARTED_KEY = "nanobot-webui.restartStartedAt";
 const SIDEBAR_WIDTH = 272;
+const TOKEN_REFRESH_MARGIN_MS = 30_000;
+const TOKEN_REFRESH_MIN_DELAY_MS = 5_000;
 type ShellView = "chat" | "settings";
 
+function bootstrapTokenExpiresAt(expiresInSeconds: number): number {
+  return Date.now() + Math.max(0, expiresInSeconds) * 1000;
+}
+
+function tokenRefreshDelayMs(expiresAt: number): number {
+  const remaining = Math.max(0, expiresAt - Date.now());
+  const margin = Math.min(
+    TOKEN_REFRESH_MARGIN_MS,
+    Math.max(1_000, remaining / 2),
+  );
+  return Math.max(TOKEN_REFRESH_MIN_DELAY_MS, remaining - margin);
+}
+
 function AuthForm({
   failed,
   onSecret,
@@ -106,6 +124,7 @@ function readSidebarOpen(): boolean {
 export default function App() {
   const { t } = useTranslation();
   const [state, setState] = useState<BootState>({ status: "loading" });
+  const bootstrapSecretRef = useRef("");
 
   const bootstrapWithSecret = useCallback(
     (secret: string) => {
@@ -117,22 +136,37 @@ export default function App() {
           if (cancelled) return;
           if (secret) saveSecret(secret);
           const url = deriveWsUrl(boot.ws_path, boot.token);
-          const client = new NanobotClient({
+          let client: NanobotClient;
+          client = new NanobotClient({
             url,
             onReauth: async () => {
               try {
-                const refreshed = await fetchBootstrap("", secret);
-                return deriveWsUrl(refreshed.ws_path, refreshed.token);
+                const refreshed = await fetchBootstrap("", bootstrapSecretRef.current);
+                const refreshedUrl = deriveWsUrl(refreshed.ws_path, refreshed.token);
+                const tokenExpiresAt = bootstrapTokenExpiresAt(refreshed.expires_in);
+                setState((current) =>
+                  current.status === "ready" && current.client === client
+                    ? {
+                        ...current,
+                        token: refreshed.token,
+                        tokenExpiresAt,
+                        modelName: refreshed.model_name ?? current.modelName,
+                      }
+                    : current,
+                );
+                return refreshedUrl;
               } catch {
                 return null;
               }
             },
           });
+          bootstrapSecretRef.current = secret;
           client.connect();
           setState({
             status: "ready",
             client,
             token: boot.token,
+            tokenExpiresAt: bootstrapTokenExpiresAt(boot.expires_in),
             modelName: boot.model_name ?? null,
           });
         } catch (e) {
@@ -152,6 +186,35 @@ export default function App() {
     [],
   );
 
+  useEffect(() => {
+    if (state.status !== "ready") return;
+    const client = state.client;
+    const timer = window.setTimeout(async () => {
+      try {
+        const boot = await fetchBootstrap("", bootstrapSecretRef.current);
+        const url = deriveWsUrl(boot.ws_path, boot.token);
+        const tokenExpiresAt = bootstrapTokenExpiresAt(boot.expires_in);
+        client.updateUrl(url);
+        setState((current) =>
+          current.status === "ready" && current.client === client
+            ? {
+                ...current,
+                token: boot.token,
+                tokenExpiresAt,
+                modelName: boot.model_name ?? current.modelName,
+              }
+            : current,
+        );
+      } catch (e) {
+        const msg = (e as Error).message;
+        if (msg.includes("HTTP 401") || msg.includes("HTTP 403")) {
+          setState({ status: "auth", failed: true });
+        }
+      }
+    }, tokenRefreshDelayMs(state.tokenExpiresAt));
+    return () => window.clearTimeout(timer);
+  }, [state]);
+
   useEffect(() => {
     const saved = loadSavedSecret();
     return bootstrapWithSecret(saved);
@@ -219,7 +282,13 @@ export default function App() {
   );
 }
 
-function Shell({ onModelNameChange, onLogout }: { onModelNameChange: (modelName: string | null) => void; onLogout: () => void }) {
+function Shell({
+  onModelNameChange,
+  onLogout,
+}: {
+  onModelNameChange: (modelName: string | null) => void;
+  onLogout: () => void;
+}) {
   const { t, i18n } = useTranslation();
   const { client } = useClient();
   const { theme, toggle } = useTheme();
@@ -362,9 +431,7 @@ function Shell({ onModelNameChange, onLogout }: { onModelNameChange: (modelName:
     });
   }, [client, t]);
 
-  const onTurnEnd = useCallback(() => {
-    void refresh();
-  }, [refresh]);
+  const onTurnEnd = useDeferredTitleRefresh(activeSession, refresh);
 
   const onConfirmDelete = useCallback(async () => {
     if (!pendingDelete) return;
@@ -386,8 +453,7 @@ function Shell({ onModelNameChange, onLogout }: { onModelNameChange: (modelName:
 
   const headerTitle = activeSession
     ? activeSession.title ||
-      activeSession.preview ||
-      t("chat.fallbackTitle", { id: activeSession.chatId.slice(0, 6) })
+      deriveTitle(activeSession.preview, t("chat.newChat"))
     : t("app.brand");
 
   useEffect(() => {
@@ -415,93 +481,95 @@ function Shell({ onModelNameChange, onLogout }: { onModelNameChange: (modelName:
   const showMainSidebar = view !== "settings";
 
   return (
-    <div className="relative flex h-full w-full overflow-hidden">
-      {/* Desktop sidebar: in normal flow, so the thread area width stays honest. */}
-      {showMainSidebar ? (
-        <aside
-          className={cn(
-            "relative z-20 hidden shrink-0 overflow-hidden lg:block",
-            "transition-[width] duration-300 ease-out",
-          )}
-          style={{ width: desktopSidebarOpen ? SIDEBAR_WIDTH : 0 }}
-        >
+    <ThemeProvider theme={theme}>
+      <div className="relative flex h-full w-full overflow-hidden">
+        {/* Desktop sidebar: in normal flow, so the thread area width stays honest. */}
+        {showMainSidebar ? (
+          <aside
+            className={cn(
+              "relative z-20 hidden shrink-0 overflow-hidden lg:block",
+              "transition-[width] duration-300 ease-out",
+            )}
+            style={{ width: desktopSidebarOpen ? SIDEBAR_WIDTH : 0 }}
+          >
+            <div
+              className={cn(
+                "absolute inset-y-0 left-0 h-full overflow-hidden bg-sidebar shadow-inner-right",
+                "transition-transform duration-300 ease-out",
+                desktopSidebarOpen ? "translate-x-0" : "-translate-x-full",
+              )}
+              style={{ width: SIDEBAR_WIDTH }}
+            >
+              <Sidebar {...sidebarProps} onCollapse={closeDesktopSidebar} />
+            </div>
+          </aside>
+        ) : null}
+
+        {showMainSidebar ? (
+          <Sheet
+            open={mobileSidebarOpen}
+            onOpenChange={(open) => setMobileSidebarOpen(open)}
+          >
+            <SheetContent
+              side="left"
+              showCloseButton={false}
+              className="p-0 lg:hidden"
+              style={{ width: SIDEBAR_WIDTH, maxWidth: SIDEBAR_WIDTH }}
+            >
+              <Sidebar {...sidebarProps} onCollapse={closeMobileSidebar} />
+            </SheetContent>
+          </Sheet>
+        ) : null}
+
+        <main className="relative flex h-full min-w-0 flex-1 flex-col">
           <div
             className={cn(
-              "absolute inset-y-0 left-0 h-full overflow-hidden bg-sidebar shadow-inner-right",
-              "transition-transform duration-300 ease-out",
-              desktopSidebarOpen ? "translate-x-0" : "-translate-x-full",
+              "absolute inset-0 flex flex-col",
+              view === "settings" && "invisible pointer-events-none",
             )}
-            style={{ width: SIDEBAR_WIDTH }}
           >
-            <Sidebar {...sidebarProps} onCollapse={closeDesktopSidebar} />
-          </div>
-        </aside>
-      ) : null}
-
-      {showMainSidebar ? (
-        <Sheet
-          open={mobileSidebarOpen}
-          onOpenChange={(open) => setMobileSidebarOpen(open)}
-        >
-          <SheetContent
-            side="left"
-            showCloseButton={false}
-            className="p-0 lg:hidden"
-            style={{ width: SIDEBAR_WIDTH, maxWidth: SIDEBAR_WIDTH }}
-          >
-            <Sidebar {...sidebarProps} onCollapse={closeMobileSidebar} />
-          </SheetContent>
-        </Sheet>
-      ) : null}
-
-      <main className="relative flex h-full min-w-0 flex-1 flex-col">
-        <div
-          className={cn(
-            "absolute inset-0 flex flex-col",
-            view === "settings" && "invisible pointer-events-none",
-          )}
-        >
-          <ThreadShell
-            session={activeSession}
-            title={headerTitle}
-            onToggleSidebar={toggleSidebar}
-            onNewChat={onNewChat}
-            onCreateChat={onCreateChat}
-            onTurnEnd={onTurnEnd}
-            theme={theme}
-            onToggleTheme={toggle}
-            hideSidebarToggleOnDesktop={desktopSidebarOpen}
-          />
-        </div>
-        {view === "settings" && (
-          <div className="absolute inset-0 flex flex-col">
-            <SettingsView
+            <ThreadShell
+              session={activeSession}
+              title={headerTitle}
+              onToggleSidebar={toggleSidebar}
+              onNewChat={onNewChat}
+              onCreateChat={onCreateChat}
+              onTurnEnd={onTurnEnd}
               theme={theme}
               onToggleTheme={toggle}
-              onBackToChat={onBackToChat}
-              onModelNameChange={onModelNameChange}
-              onLogout={onLogout}
-              onRestart={onRestart}
-              isRestarting={isRestarting}
+              hideSidebarToggleOnDesktop={desktopSidebarOpen}
             />
           </div>
-        )}
-      </main>
+          {view === "settings" && (
+            <div className="absolute inset-0 flex flex-col">
+              <SettingsView
+                theme={theme}
+                onToggleTheme={toggle}
+                onBackToChat={onBackToChat}
+                onModelNameChange={onModelNameChange}
+                onLogout={onLogout}
+                onRestart={onRestart}
+                isRestarting={isRestarting}
+              />
+            </div>
+          )}
+        </main>
 
-      <DeleteConfirm
-        open={!!pendingDelete}
-        title={pendingDelete?.label ?? ""}
-        onCancel={() => setPendingDelete(null)}
-        onConfirm={onConfirmDelete}
-      />
-      {restartToast ? (
-        <div
-          role="status"
-          className="fixed left-1/2 top-4 z-50 -translate-x-1/2 rounded-full border border-border/70 bg-popover px-4 py-2 text-sm font-medium text-popover-foreground shadow-lg"
-        >
-          {restartToast}
-        </div>
-      ) : null}
-    </div>
+        <DeleteConfirm
+          open={!!pendingDelete}
+          title={pendingDelete?.label ?? ""}
+          onCancel={() => setPendingDelete(null)}
+          onConfirm={onConfirmDelete}
+        />
+        {restartToast ? (
+          <div
+            role="status"
+            className="fixed left-1/2 top-4 z-50 -translate-x-1/2 rounded-full border border-border/70 bg-popover px-4 py-2 text-sm font-medium text-popover-foreground shadow-lg"
+          >
+            {restartToast}
+          </div>
+        ) : null}
+      </div>
+    </ThemeProvider>
   );
 }

webui/src/components/ChatList.tsx

@@ -7,6 +7,7 @@ import {
   DropdownMenuItem,
   DropdownMenuTrigger,
 } from "@/components/ui/dropdown-menu";
+import { deriveTitle } from "@/lib/format";
 import { cn } from "@/lib/utils";
 import type { ChatSummary } from "@/lib/types";
 
@@ -64,8 +65,11 @@ export function ChatList({
                 const fallbackTitle = t("chat.fallbackTitle", {
                   id: s.chatId.slice(0, 6),
                 });
-                const rawLabel = (s.title || s.preview)?.trim();
-                const title = rawLabel || fallbackTitle;
+                const generatedTitle = s.title?.trim() || "";
+                const title =
+                  generatedTitle || deriveTitle(s.preview, t("chat.newChat"));
+                const tooltipTitle =
+                  generatedTitle || deriveTitle(s.preview, fallbackTitle);
                 return (
                   <li key={s.key} className="min-w-0">
                     <div
@@ -79,7 +83,7 @@ export function ChatList({
                       <button
                         type="button"
                         onClick={() => onSelect(s.key)}
-                        title={rawLabel || fallbackTitle}
+                        title={tooltipTitle}
                         className="min-w-0 flex-1 overflow-hidden py-1.5 text-left"
                       >
                         <span className="block w-full truncate font-medium leading-5">{title}</span>

webui/src/components/CodeBlock.tsx

@@ -1,44 +1,75 @@
-import { useCallback, useEffect, useState } from "react";
+import { Suspense, lazy, useCallback, useState } from "react";
 import { Check, Copy } from "lucide-react";
 import { useTranslation } from "react-i18next";
-import { Prism as SyntaxHighlighter } from "react-syntax-highlighter";
-import {
-  oneDark,
-  oneLight,
-} from "react-syntax-highlighter/dist/esm/styles/prism";
 
+import { useThemeValue } from "@/hooks/useTheme";
 import { cn } from "@/lib/utils";
 
 interface CodeBlockProps {
   language?: string;
   code: string;
   className?: string;
+  highlight?: boolean;
 }
 
-/** Read dark mode straight from the DOM — stays in sync with Tailwind's `dark:`. */
-function useIsDark() {
-  const [isDark, setIsDark] = useState(() =>
-    typeof document !== "undefined"
-      ? document.documentElement.classList.contains("dark")
-      : true,
+interface HighlightedCodeProps {
+  language?: string;
+  code: string;
+  isDark: boolean;
+}
+
+const LazyHighlightedCode = lazy(async () => {
+  const [
+    { default: SyntaxHighlighter },
+    { default: oneDark },
+    { default: oneLight },
+  ] = await Promise.all([
+    import("react-syntax-highlighter/dist/esm/prism-async-light"),
+    import("react-syntax-highlighter/dist/esm/styles/prism/one-dark"),
+    import("react-syntax-highlighter/dist/esm/styles/prism/one-light"),
+  ]);
+
+  return {
+    default({ language, code, isDark }: HighlightedCodeProps) {
+      return (
+        <SyntaxHighlighter
+          language={language}
+          style={isDark ? oneDark : oneLight}
+          customStyle={{
+            margin: 0,
+            padding: "1rem",
+            fontSize: "0.875rem",
+            lineHeight: 1.6,
+          }}
+          PreTag="pre"
+          wrapLongLines
+        >
+          {code}
+        </SyntaxHighlighter>
+      );
+    },
+  };
+});
+
+function PlainCodeFallback({ code }: { code: string }) {
+  return (
+    <pre
+      className="m-0 overflow-x-auto whitespace-pre-wrap p-4 font-mono text-sm leading-[1.6]"
+    >
+      <code>{code}</code>
+    </pre>
   );
-
-  useEffect(() => {
-    const el = document.documentElement;
-    const observer = new MutationObserver(() => {
-      setIsDark(el.classList.contains("dark"));
-    });
-    observer.observe(el, { attributeFilter: ["class"] });
-    return () => observer.disconnect();
-  }, []);
-
-  return isDark;
 }
 
-export function CodeBlock({ language, code, className }: CodeBlockProps) {
+export function CodeBlock({
+  language,
+  code,
+  className,
+  highlight = true,
+}: CodeBlockProps) {
   const { t } = useTranslation();
   const [copied, setCopied] = useState(false);
-  const isDark = useIsDark();
+  const isDark = useThemeValue() === "dark";
 
   const onCopy = useCallback(() => {
     if (!navigator.clipboard) return;
@@ -86,20 +117,13 @@ export function CodeBlock({ language, code, className }: CodeBlockProps) {
           <span>{copied ? t("code.copied") : t("code.copy")}</span>
         </button>
       </div>
-      <SyntaxHighlighter
-        language={language}
-        style={isDark ? oneDark : oneLight}
-        customStyle={{
-          margin: 0,
-          padding: "1rem",
-          fontSize: "0.875rem",
-          lineHeight: 1.6,
-        }}
-        PreTag="pre"
-        wrapLongLines
-      >
-        {code}
-      </SyntaxHighlighter>
+      {highlight ? (
+        <Suspense fallback={<PlainCodeFallback code={code} />}>
+          <LazyHighlightedCode language={language} code={code} isDark={isDark} />
+        </Suspense>
+      ) : (
+        <PlainCodeFallback code={code} />
+      )}
     </div>
   );
 }

webui/src/components/ConnectionBadge.tsx

@@ -36,21 +36,25 @@ export function ConnectionBadge() {
     status === "connecting" ||
     status === "reconnecting" ||
     status === "error";
+  const label = t(`connection.${status}`);
   return (
     <span
       className={cn(
-        "inline-flex min-w-0 items-center gap-1.5 rounded-md px-1.5 py-1 text-[11px] font-medium transition-colors",
+        "inline-flex h-8 w-8 shrink-0 items-center justify-center rounded-full transition-colors",
+        "text-muted-foreground/70 hover:bg-sidebar-accent/65",
         meta.color,
       )}
       aria-live="polite"
+      role="status"
+      title={label}
     >
-      <span className="relative flex h-1.5 w-1.5" aria-hidden>
+      <span className="relative flex h-2 w-2" aria-hidden>
         {pulsing && (
           <span className="absolute inline-flex h-full w-full animate-ping rounded-full bg-current opacity-75" />
         )}
-        <span className="relative inline-flex h-1.5 w-1.5 rounded-full bg-current" />
+        <span className="relative inline-flex h-2 w-2 rounded-full bg-current" />
       </span>
-      {t(`connection.${status}`)}
+      <span className="sr-only">{label}</span>
     </span>
   );
 }

webui/src/components/FileReferenceChip.tsx

@@ -0,0 +1,220 @@
+import {
+  Tooltip,
+  TooltipContent,
+  TooltipProvider,
+  TooltipTrigger,
+} from "@/components/ui/tooltip";
+import { cn } from "@/lib/utils";
+
+type FileReferenceKind =
+  | "default"
+  | "css"
+  | "html"
+  | "json"
+  | "markdown"
+  | "notebook"
+  | "python"
+  | "react"
+  | "typescript";
+
+interface FileReferenceChipProps {
+  path: string;
+  display?: "name" | "path";
+  active?: boolean;
+  className?: string;
+  textClassName?: string;
+  testId?: string;
+}
+
+export function FileReferenceChip({
+  path,
+  display = "name",
+  active = false,
+  className,
+  textClassName,
+  testId = "inline-file-path",
+}: FileReferenceChipProps) {
+  const { name } = splitFilePath(path);
+  const kind = fileKindForPath(path);
+  const displayText = display === "path" ? path.replace(/\\/g, "/") : name;
+  return (
+    <TooltipProvider delayDuration={500} skipDelayDuration={100}>
+      <Tooltip>
+        <TooltipTrigger asChild>
+          <span
+            className={cn("not-prose inline-flex max-w-full align-[0.14em]", className)}
+          >
+            <span
+              data-testid={testId}
+              aria-label={path}
+              className={cn(
+                "inline-flex max-w-full items-center gap-1 font-medium leading-[1.1]",
+                "text-sky-600 transition-colors hover:text-sky-700",
+                "dark:text-sky-300 dark:hover:text-sky-200",
+              )}
+            >
+              <FileReferenceIcon kind={kind} />
+              <span
+                data-sheen-text={active ? displayText : undefined}
+                className={cn(
+                  "min-w-0 truncate",
+                  active && "streaming-text-sheen",
+                  textClassName,
+                )}
+              >
+                {displayText}
+              </span>
+            </span>
+          </span>
+        </TooltipTrigger>
+        <TooltipContent
+          side="top"
+          align="center"
+          sideOffset={8}
+          collisionPadding={12}
+          className={cn(
+            "max-w-[min(38rem,calc(100vw-2rem))] rounded-[10px]",
+            "border-border/60 bg-popover/95 px-2.5 py-1.5",
+            "break-all font-mono text-[11px] leading-snug text-popover-foreground",
+            "shadow-lg backdrop-blur",
+          )}
+        >
+          {path}
+        </TooltipContent>
+      </Tooltip>
+    </TooltipProvider>
+  );
+}
+
+export function isLikelyFilePath(value: string): boolean {
+  const raw = value.trim();
+  if (!raw || raw.includes("\n")) return false;
+  if (/^[a-z][a-z0-9+.-]*:\/\//i.test(raw)) return false;
+  if (!/[\\/]/.test(raw) && !/^(dockerfile|makefile|readme|package-lock\.json)$/i.test(raw)) {
+    return false;
+  }
+  const normalized = raw.replace(/\\/g, "/");
+  const name = normalized.split("/").filter(Boolean).pop() ?? normalized;
+  if (!name || name === "." || name === "..") return false;
+  if (/^(dockerfile|makefile|readme|package-lock\.json)$/i.test(name)) return true;
+  return /\.[a-z0-9][a-z0-9_-]{0,12}$/i.test(name);
+}
+
+function splitFilePath(path: string): { directory: string; name: string } {
+  const normalized = path.replace(/\\/g, "/");
+  const slash = normalized.lastIndexOf("/");
+  if (slash < 0) return { directory: "", name: path };
+  return {
+    directory: normalized.slice(0, slash + 1),
+    name: normalized.slice(slash + 1) || normalized,
+  };
+}
+
+function fileKindForPath(path: string): FileReferenceKind {
+  const normalized = path.toLowerCase();
+  const name = normalized.split(/[\\/]/).pop() ?? normalized;
+  const ext = name.includes(".") ? name.split(".").pop() ?? "" : "";
+  if (name === "dockerfile") {
+    return "default";
+  }
+  switch (ext) {
+    case "py":
+    case "pyi":
+      return "python";
+    case "jsx":
+    case "tsx":
+      return "react";
+    case "ts":
+      return "typescript";
+    case "html":
+    case "htm":
+      return "html";
+    case "css":
+    case "scss":
+    case "sass":
+      return "css";
+    case "json":
+    case "jsonl":
+      return "json";
+    case "md":
+    case "mdx":
+      return "markdown";
+    case "ipynb":
+      return "notebook";
+    default:
+      return "default";
+  }
+}
+
+function FileReferenceIcon({ kind }: { kind: FileReferenceKind }) {
+  if (kind === "react") {
+    return (
+      <svg
+        aria-hidden
+        className="h-[0.98em] w-[0.98em] shrink-0 text-sky-500 dark:text-sky-300"
+        viewBox="0 0 24 24"
+        fill="none"
+        stroke="currentColor"
+        strokeWidth="1.6"
+        strokeLinecap="round"
+        strokeLinejoin="round"
+      >
+        <circle cx="12" cy="12" r="1.9" fill="currentColor" stroke="none" />
+        <ellipse cx="12" cy="12" rx="9" ry="3.7" />
+        <ellipse cx="12" cy="12" rx="9" ry="3.7" transform="rotate(60 12 12)" />
+        <ellipse cx="12" cy="12" rx="9" ry="3.7" transform="rotate(120 12 12)" />
+      </svg>
+    );
+  }
+  if (kind === "default") {
+    return (
+      <svg
+        aria-hidden
+        className="h-[0.98em] w-[0.98em] shrink-0 text-sky-500 dark:text-sky-300"
+        viewBox="0 0 24 24"
+        fill="none"
+        stroke="currentColor"
+        strokeWidth="1.9"
+        strokeLinecap="round"
+        strokeLinejoin="round"
+      >
+        <path d="M14 2H7a2 2 0 0 0-2 2v16a2 2 0 0 0 2 2h10a2 2 0 0 0 2-2V7z" />
+        <path d="M14 2v5h5" />
+      </svg>
+    );
+  }
+  const label = fileKindLabel(kind);
+  return (
+    <span
+      aria-hidden
+      className={cn(
+        "inline-flex h-[1.05em] min-w-[1.05em] shrink-0 items-center justify-center",
+        "rounded-[4px] bg-sky-500/12 px-[0.22em] text-[0.58em] font-bold uppercase leading-none",
+        "text-sky-600 dark:bg-sky-400/15 dark:text-sky-300",
+      )}
+    >
+      {label}
+    </span>
+  );
+}
+
+function fileKindLabel(kind: FileReferenceKind): string {
+  switch (kind) {
+    case "css":
+      return "#";
+    case "html":
+      return "H";
+    case "json":
+      return "{}";
+    case "markdown":
+      return "M";
+    case "notebook":
+      return "N";
+    case "python":
+      return "PY";
+    case "typescript":
+      return "TS";
+    default:
+      return "";
+  }
+}

webui/src/components/MarkdownText.tsx

@@ -1,15 +1,46 @@
-import { Suspense, lazy } from "react";
+import {
+  Suspense,
+  lazy,
+  memo,
+  startTransition,
+  useCallback,
+  useEffect,
+  useLayoutEffect,
+  useRef,
+  useState,
+} from "react";
 
 import { cn } from "@/lib/utils";
 
 interface MarkdownTextProps {
   children: string;
   className?: string;
+  streaming?: boolean;
 }
 
 const loadMarkdownRenderer = () => import("@/components/MarkdownTextRenderer");
 const LazyMarkdownRenderer = lazy(loadMarkdownRenderer);
 
+const MemoizedMarkdownRenderer = memo(function MemoizedMarkdownRenderer({
+  source,
+  className,
+  highlightCode,
+}: {
+  source: string;
+  className?: string;
+  highlightCode: boolean;
+}) {
+  return (
+    <LazyMarkdownRenderer className={className} highlightCode={highlightCode}>
+      {source}
+    </LazyMarkdownRenderer>
+  );
+});
+
+const SHORT_STREAM_COMMIT_MS = 80;
+const MEDIUM_STREAM_COMMIT_MS = 140;
+const LONG_STREAM_COMMIT_MS = 220;
+
 export function preloadMarkdownText(): void {
   void loadMarkdownRenderer();
 }
@@ -19,7 +50,18 @@ export function preloadMarkdownText(): void {
  * ``remark-math`` / ``rehype-katex``, and fenced code blocks delegated to
  * ``CodeBlock`` for copy-to-clipboard and syntax highlighting.
  */
-export function MarkdownText({ children, className }: MarkdownTextProps) {
+export function MarkdownText({
+  children,
+  className,
+  streaming = false,
+}: MarkdownTextProps) {
+  const renderedSource = useStreamingMarkdownSource(children, streaming);
+  const highlightCode = !streaming && renderedSource === children;
+
+  useEffect(() => {
+    if (streaming) preloadMarkdownText();
+  }, [streaming]);
+
   return (
     <Suspense
       fallback={
@@ -29,11 +71,73 @@ export function MarkdownText({ children, className }: MarkdownTextProps) {
             className,
           )}
         >
-          {children}
+          {renderedSource}
         </div>
       }
     >
-      <LazyMarkdownRenderer className={className}>{children}</LazyMarkdownRenderer>
+      <MemoizedMarkdownRenderer
+        source={renderedSource}
+        className={className}
+        highlightCode={highlightCode}
+      />
     </Suspense>
   );
 }
+
+function useStreamingMarkdownSource(source: string, streaming: boolean): string {
+  const [renderedSource, setRenderedSource] = useState(source);
+  const latestSourceRef = useRef(source);
+  const renderedSourceRef = useRef(source);
+  const timerRef = useRef<number | null>(null);
+
+  const clearPendingCommit = useCallback(() => {
+    if (timerRef.current !== null) {
+      window.clearTimeout(timerRef.current);
+      timerRef.current = null;
+    }
+  }, []);
+
+  const commitSource = useCallback((next: string, urgent: boolean) => {
+    if (renderedSourceRef.current === next) return;
+    renderedSourceRef.current = next;
+    if (urgent) {
+      setRenderedSource(next);
+      return;
+    }
+    startTransition(() => setRenderedSource(next));
+  }, []);
+
+  const scheduleCommit = useCallback(() => {
+    if (timerRef.current !== null) return;
+    timerRef.current = window.setTimeout(() => {
+      timerRef.current = null;
+      commitSource(latestSourceRef.current, false);
+    }, streamingCommitDelay(latestSourceRef.current.length));
+  }, [commitSource]);
+
+  latestSourceRef.current = source;
+
+  useLayoutEffect(() => {
+    latestSourceRef.current = source;
+    if (!streaming) {
+      clearPendingCommit();
+      commitSource(source, true);
+    }
+  }, [clearPendingCommit, commitSource, source, streaming]);
+
+  useEffect(() => {
+    latestSourceRef.current = source;
+    if (!streaming) return;
+    scheduleCommit();
+  }, [scheduleCommit, source, streaming]);
+
+  useEffect(() => clearPendingCommit, [clearPendingCommit]);
+
+  return renderedSource;
+}
+
+function streamingCommitDelay(length: number): number {
+  if (length > 24_000) return LONG_STREAM_COMMIT_MS;
+  if (length > 8_000) return MEDIUM_STREAM_COMMIT_MS;
+  return SHORT_STREAM_COMMIT_MS;
+}

webui/src/components/MarkdownTextRenderer.tsx

@@ -1,10 +1,13 @@
-import { Children, isValidElement } from "react";
+import { Children, isValidElement, useMemo } from "react";
+import type { Components } from "react-markdown";
 import ReactMarkdown from "react-markdown";
 import rehypeKatex from "rehype-katex";
+import remarkBreaks from "remark-breaks";
 import remarkGfm from "remark-gfm";
 import remarkMath from "remark-math";
 
 import { CodeBlock } from "@/components/CodeBlock";
+import { FileReferenceChip, isLikelyFilePath } from "@/components/FileReferenceChip";
 import { cn } from "@/lib/utils";
 
 import "katex/dist/katex.min.css";
@@ -12,8 +15,12 @@ import "katex/dist/katex.min.css";
 interface MarkdownTextRendererProps {
   children: string;
   className?: string;
+  highlightCode?: boolean;
 }
 
+const remarkPlugins = [remarkBreaks, remarkGfm, remarkMath];
+const rehypePlugins = [rehypeKatex];
+
 /**
  * Heavy markdown stack (GFM, math, KaTeX, syntax highlighting) kept in a
  * separate chunk so the app shell can paint sooner on refresh.
@@ -21,7 +28,91 @@ interface MarkdownTextRendererProps {
 export default function MarkdownTextRenderer({
   children,
   className,
+  highlightCode = true,
 }: MarkdownTextRendererProps) {
+  const components = useMemo<Components>(
+    () => ({
+      code({ className: cls, children: kids, ...props }) {
+        const match = /language-(\w+)/.exec(cls || "");
+        if (match) {
+          const code = String(kids).replace(/\n$/, "");
+          return (
+            <CodeBlock
+              language={match[1]}
+              code={code}
+              className="my-3"
+              highlight={highlightCode}
+            />
+          );
+        }
+        const raw = String(kids).replace(/\n$/, "");
+        if (isLikelyFilePath(raw)) {
+          return <FileReferenceChip path={raw} />;
+        }
+        /** Plain fenced ``` blocks (no language) & wide one-liners: block monospace, not inline pill. */
+        const widePlainBlock = raw.includes("\n") || raw.length > 120;
+        if (widePlainBlock) {
+          return (
+            <code
+              className={cn(
+                "block min-w-0 whitespace-pre bg-transparent p-0 font-mono text-[0.8125rem]",
+                "leading-snug text-inherit",
+                cls,
+              )}
+              {...props}
+            >
+              {kids}
+            </code>
+          );
+        }
+        return (
+          <code
+            className={cn(
+              "rounded bg-muted px-1 py-0.5 font-mono text-[0.85em]",
+              cls,
+            )}
+            {...props}
+          >
+            {kids}
+          </code>
+        );
+      },
+      pre({ children: markdownChildren }) {
+        const kids = Children.toArray(markdownChildren);
+        const lone = kids.length === 1 ? kids[0] : null;
+        /** Highlighted fences render ``CodeBlock`` (block shell); skip invalid ``<pre><div>``. */
+        if (lone != null && isValidElement(lone) && lone.type === CodeBlock) {
+          return <>{markdownChildren}</>;
+        }
+        return (
+          <pre
+            className={cn(
+              "my-3 overflow-x-auto rounded-lg border border-border/60 bg-muted/35",
+              "p-3 font-mono text-[0.8125rem] leading-snug text-foreground/90",
+              "whitespace-pre [overflow-wrap:normal]",
+            )}
+          >
+            {markdownChildren}
+          </pre>
+        );
+      },
+      a({ href, children: markdownChildren, ...props }) {
+        return (
+          <a
+            href={href}
+            target="_blank"
+            rel="noreferrer noopener"
+            className="text-primary underline underline-offset-2 hover:opacity-80"
+            {...props}
+          >
+            {markdownChildren}
+          </a>
+        );
+      },
+    }),
+    [highlightCode],
+  );
+
   return (
     <div
       className={cn(
@@ -42,77 +133,9 @@ export default function MarkdownTextRenderer({
       style={{ lineHeight: "var(--cjk-line-height)" }}
     >
       <ReactMarkdown
-        remarkPlugins={[remarkGfm, remarkMath]}
-        rehypePlugins={[rehypeKatex]}
-        components={{
-          code({ className: cls, children: kids, ...props }) {
-            const match = /language-(\w+)/.exec(cls || "");
-            if (match) {
-              const code = String(kids).replace(/\n$/, "");
-              return <CodeBlock language={match[1]} code={code} className="my-3" />;
-            }
-            const raw = String(kids).replace(/\n$/, "");
-            /** Plain fenced ``` blocks (no language) & wide one-liners: block monospace, not inline pill. */
-            const widePlainBlock = raw.includes("\n") || raw.length > 120;
-            if (widePlainBlock) {
-              return (
-                <code
-                  className={cn(
-                    "block min-w-0 whitespace-pre bg-transparent p-0 font-mono text-[0.8125rem]",
-                    "leading-snug text-inherit",
-                    cls,
-                  )}
-                  {...props}
-                >
-                  {kids}
-                </code>
-              );
-            }
-            return (
-              <code
-                className={cn(
-                  "rounded bg-muted px-1 py-0.5 font-mono text-[0.85em]",
-                  cls,
-                )}
-                {...props}
-              >
-                {kids}
-              </code>
-            );
-          },
-          pre({ children: markdownChildren }) {
-            const kids = Children.toArray(markdownChildren);
-            const lone = kids.length === 1 ? kids[0] : null;
-            /** Highlighted fences render ``CodeBlock`` (block shell); skip invalid ``<pre><div>``. */
-            if (lone != null && isValidElement(lone) && lone.type === CodeBlock) {
-              return <>{markdownChildren}</>;
-            }
-            return (
-              <pre
-                className={cn(
-                  "my-3 overflow-x-auto rounded-lg border border-border/60 bg-muted/35",
-                  "p-3 font-mono text-[0.8125rem] leading-snug text-foreground/90",
-                  "whitespace-pre [overflow-wrap:normal]",
-                )}
-              >
-                {markdownChildren}
-              </pre>
-            );
-          },
-          a({ href, children: markdownChildren, ...props }) {
-            return (
-              <a
-                href={href}
-                target="_blank"
-                rel="noreferrer noopener"
-                className="text-primary underline underline-offset-2 hover:opacity-80"
-                {...props}
-              >
-                {markdownChildren}
-              </a>
-            );
-          },
-        }}
+        remarkPlugins={remarkPlugins}
+        rehypePlugins={rehypePlugins}
+        components={components}
       >
         {children}
       </ReactMarkdown>

webui/src/components/MessageBubble.tsx

@@ -1,6 +1,5 @@
 import {
   useCallback,
-  useDeferredValue,
   useEffect,
   useRef,
   useState,
@@ -120,7 +119,7 @@ export function MessageBubble({
         <TypingDots />
       ) : empty && message.isStreaming ? null : (
         <>
-          <MarkdownText>{message.content}</MarkdownText>
+          <MarkdownText streaming={!!message.isStreaming}>{message.content}</MarkdownText>
           {media.length > 0 ? <MessageMedia media={media} align="left" /> : null}
           {showAssistantFooterRow ? (
             <div className="mt-2 flex min-h-8 flex-wrap items-center gap-x-2 gap-y-1 text-muted-foreground">
@@ -167,10 +166,15 @@ function MessageMedia({
   align: "left" | "right";
 }) {
   if (media.length === 0) return null;
-  const images = media
-    .filter((item) => item.kind === "image")
-    .map(({ url, name }) => ({ url, name }));
-  const nonImages = media.filter((item) => item.kind !== "image");
+  const images: UIImage[] = [];
+  const nonImages: UIMediaAttachment[] = [];
+  for (const item of media) {
+    if (item.kind === "image") {
+      images.push({ url: item.url, name: item.name });
+    } else {
+      nonImages.push(item);
+    }
+  }
 
   return (
     <div
@@ -276,13 +280,14 @@ function UserImages({
   const { t } = useTranslation();
   // Only real-URL images can open in the lightbox; historical-replay
   // placeholders (no URL) have nothing to zoom into.
-  const viewable = images
-    .map((img, i) => ({ img, i }))
-    .filter(({ img }) => typeof img.url === "string" && img.url.length > 0);
-  const viewableImages = viewable.map(({ img }) => img);
-  const originalToViewable = new Map<number, number>(
-    viewable.map(({ i }, v) => [i, v]),
-  );
+  const viewableImages: UIImage[] = [];
+  const originalToViewable = new Map<number, number>();
+  for (let i = 0; i < images.length; i += 1) {
+    const img = images[i];
+    if (typeof img.url !== "string" || img.url.length === 0) continue;
+    originalToViewable.set(i, viewableImages.length);
+    viewableImages.push(img);
+  }
 
   const [lightboxIndex, setLightboxIndex] = useState<number | null>(null);
 
@@ -416,7 +421,7 @@ function Dot({ delay }: { delay: string }) {
   );
 }
 
-/** L→R sheen overlay on label text; base copy stays solid ``text-muted-foreground``. */
+/** L→R sheen on the glyphs themselves; inactive labels stay solid muted text. */
 export function StreamingLabelSheen({
   children,
   active,
@@ -426,21 +431,21 @@ export function StreamingLabelSheen({
   active: boolean;
   className?: string;
 }) {
+  const sheenText =
+    typeof children === "string" || typeof children === "number"
+      ? String(children)
+      : undefined;
   return (
-    <span className={cn("relative block min-w-0 py-px", className)}>
+    <span className={cn("block min-w-0 overflow-hidden py-px", className)}>
       <span
+        data-sheen-text={active ? sheenText : undefined}
         className={cn(
-          "relative z-0 block font-medium leading-normal text-muted-foreground",
-          !active && "truncate",
+          "block w-fit max-w-full truncate font-medium leading-normal",
+          active ? "streaming-text-sheen" : "text-muted-foreground",
         )}
       >
         {children}
       </span>
-      {active ? (
-        <span className="reasoning-sheen-track" aria-hidden dir="ltr">
-          <span className="reasoning-sheen-stripe" />
-        </span>
-      ) : null}
     </span>
   );
 }
@@ -474,8 +479,6 @@ export function ReasoningBubble({
   embeddedInCluster = false,
 }: ReasoningBubbleProps) {
   const { t } = useTranslation();
-  const deferredText = useDeferredValue(text);
-  const markdownSource = streaming ? deferredText : text;
   const [userToggled, setUserToggled] = useState(false);
   const [openLocal, setOpenLocal] = useState(true);
   const open = userToggled ? openLocal : streaming;
@@ -531,6 +534,7 @@ export function ReasoningBubble({
           )}
         >
           <MarkdownText
+            streaming={streaming}
             className={cn(
               "text-[12.5px] italic text-muted-foreground/88",
               "prose-p:my-1.5 prose-li:my-0.5",
@@ -541,7 +545,7 @@ export function ReasoningBubble({
               "prose-code:text-[0.92em]",
             )}
           >
-            {markdownSource}
+            {text}
           </MarkdownText>
         </div>
       )}

webui/src/components/Sidebar.tsx

@@ -117,12 +117,12 @@ export function Sidebar(props: SidebarProps) {
         />
       </div>
       <Separator className="bg-sidebar-border/50" />
-      <div className="space-y-1 px-2.5 py-2.5 text-xs">
+      <div className="flex items-center gap-1 px-2.5 py-2.5 text-xs">
         <Button
           type="button"
           variant="ghost"
           onClick={props.onOpenSettings}
-          className="h-8 w-full justify-start gap-2 rounded-full px-2.5 text-[12.5px] font-medium text-sidebar-foreground/85 hover:bg-sidebar-accent/75 hover:text-sidebar-foreground"
+          className="h-8 min-w-0 flex-1 justify-start gap-2 rounded-full px-2.5 text-[12.5px] font-medium text-sidebar-foreground/85 hover:bg-sidebar-accent/75 hover:text-sidebar-foreground"
         >
           <Settings className="h-3.5 w-3.5" aria-hidden />
           {t("sidebar.settings")}

webui/src/components/settings/SettingsView.tsx

@@ -52,6 +52,13 @@ import type { SettingsPayload, WebSearchSettingsUpdate } from "@/lib/types";
 type SettingsSectionKey = "general" | "byok";
 type ByokPaneKey = "llm" | "web-search";
 
+const LOCAL_UNCONFIGURED_PROVIDER_ORDER = new Map(
+  ["vllm", "ollama", "lm_studio", "atomic_chat", "ovms"].map((name, index) => [
+    name,
+    index,
+  ]),
+);
+
 interface SettingsViewProps {
   theme: "light" | "dark";
   onToggleTheme: () => void;
@@ -176,7 +183,8 @@ export function SettingsView({
     if (!provider) return;
     const providerForm = providerForms[providerName] ?? { apiKey: "", apiBase: "" };
     const apiKey = providerForm.apiKey.trim();
-    if (!provider.configured && !apiKey) {
+    const apiKeyRequired = provider.api_key_required ?? true;
+    if (!provider.configured && apiKeyRequired && !apiKey) {
       setError(t("settings.byok.apiKeyRequired"));
       return;
     }
@@ -917,7 +925,10 @@ function ByokSettings({
   const [activePane, setActivePane] = useState<ByokPaneKey>("llm");
   const [showAllUnconfigured, setShowAllUnconfigured] = useState(false);
   const configuredProviders = settings.providers.filter((provider) => provider.configured);
-  const unconfiguredProviders = settings.providers.filter((provider) => !provider.configured);
+  const unconfiguredProviders = useMemo(
+    () => orderUnconfiguredProviders(settings.providers.filter((provider) => !provider.configured)),
+    [settings.providers],
+  );
   const initialUnconfiguredCount = 6;
   const visibleUnconfiguredProviders = showAllUnconfigured
     ? unconfiguredProviders
@@ -935,6 +946,12 @@ function ByokSettings({
     const saving = providerSaving === provider.name;
     const keyVisible = !!visibleProviderKeys[provider.name];
     const editingKey = !provider.configured || !!editingProviderKeys[provider.name];
+    const apiKeyRequired = provider.api_key_required ?? true;
+    const apiKey = form.apiKey.trim();
+    const apiBase = form.apiBase.trim();
+    const missingRequiredApiKey = apiKeyRequired && !provider.configured && !apiKey;
+    const missingOptionalCredential =
+      !apiKeyRequired && !provider.configured && !apiKey && !apiBase;
     return (
       <div
         key={provider.name}
@@ -1045,7 +1062,7 @@ function ByokSettings({
                 size="sm"
                 variant="outline"
                 onClick={() => onSaveProvider(provider.name)}
-                disabled={saving || (!provider.configured && !form.apiKey.trim())}
+                disabled={saving || missingRequiredApiKey || missingOptionalCredential}
                 className="rounded-full"
               >
                 {saving ? t("settings.actions.saving") : t("settings.actions.save")}
@@ -1188,6 +1205,25 @@ function ByokEmptyState({ children }: { children: ReactNode }) {
   );
 }
 
+function orderUnconfiguredProviders(
+  providers: SettingsPayload["providers"],
+): SettingsPayload["providers"] {
+  return providers
+    .map((provider, index) => ({ provider, index }))
+    .sort((left, right) => {
+      const rank = providerVisibilityRank(left.provider) - providerVisibilityRank(right.provider);
+      return rank || left.index - right.index;
+    })
+    .map(({ provider }) => provider);
+}
+
+function providerVisibilityRank(provider: SettingsPayload["providers"][number]): number {
+  const localRank = LOCAL_UNCONFIGURED_PROVIDER_ORDER.get(provider.name);
+  if (localRank !== undefined) return localRank;
+  if ((provider.api_key_required ?? true) === false) return 100;
+  return 200;
+}
+
 const PROVIDER_ICONS: Record<string, LucideIcon> = {
   custom: Hexagon,
   openrouter: Sparkles,
@@ -1212,6 +1248,12 @@ const PROVIDER_ICONS: Record<string, LucideIcon> = {
   qianfan: Database,
   azure_openai: Cloud,
   bedrock: Database,
+  vllm: Cpu,
+  ollama: Cpu,
+  lm_studio: Cpu,
+  atomic_chat: Cpu,
+  ovms: Cpu,
+  nvidia: Zap,
 };
 
 function ProviderIcon({ provider }: { provider: string }) {

webui/src/components/thread/AgentActivityCluster.tsx

@@ -1,13 +1,15 @@
-import { useState } from "react";
-import { ChevronRight, Layers } from "lucide-react";
+import { useCallback, useEffect, useLayoutEffect, useMemo, useRef, useState } from "react";
+import { AlertCircle, ChevronRight, Layers } from "lucide-react";
 import { useTranslation } from "react-i18next";
 
+import { FileReferenceChip } from "@/components/FileReferenceChip";
 import { ReasoningBubble, StreamingLabelSheen, TraceGroup } from "@/components/MessageBubble";
 import { cn } from "@/lib/utils";
-import type { UIMessage } from "@/lib/types";
+import type { UIFileEdit, UIMessage } from "@/lib/types";
 
 /** Scrollport height for the Cursor-style “live trace” strip (tailwind spacing). */
 const CLUSTER_SCROLL_MAX_CLASS = "max-h-52";
+const ACTIVITY_SCROLL_NEAR_BOTTOM_PX = 24;
 
 export function isReasoningOnlyAssistant(m: UIMessage): boolean {
   if (m.role !== "assistant" || m.kind === "trace") return false;
@@ -19,14 +21,70 @@ export function isAgentActivityMember(m: UIMessage): boolean {
   return isReasoningOnlyAssistant(m) || m.kind === "trace";
 }
 
-function countToolCalls(messages: UIMessage[]): number {
-  let n = 0;
+interface ActivityCounts {
+  reasoningSteps: number;
+  toolCalls: number;
+  fileCount: number;
+  added: number;
+  deleted: number;
+  hasEditingFiles: boolean;
+  hasFailedFiles: boolean;
+  primaryFilePath?: string;
+}
+
+interface FileEditSummary {
+  key: string;
+  path: string;
+  added: number;
+  deleted: number;
+  approximate: boolean;
+  binary: boolean;
+  status: UIFileEdit["status"];
+  error?: string;
+}
+
+function countActivity(messages: UIMessage[], fileEdits: FileEditSummary[]): ActivityCounts {
+  let reasoningSteps = 0;
+  let toolCalls = 0;
   for (const m of messages) {
-    if (m.kind !== "trace") continue;
-    const lines = m.traces?.length ?? (m.content.trim() ? 1 : 0);
-    n += Math.max(lines, 1);
+    if (isReasoningOnlyAssistant(m)) {
+      reasoningSteps += 1;
+      continue;
+    }
+    if (m.kind === "trace") {
+      const lines = m.traces?.length ?? (m.content.trim() ? 1 : 0);
+      toolCalls += lines;
+    }
   }
-  return n;
+  let added = 0;
+  let deleted = 0;
+  let hasEditingFiles = false;
+  let failedFileCount = 0;
+  let primaryFilePath: string | undefined;
+  for (const edit of fileEdits) {
+    primaryFilePath = edit.path;
+    if (edit.status === "editing") {
+      hasEditingFiles = true;
+    }
+    if (edit.status === "error") {
+      failedFileCount += 1;
+    }
+    if (edit.status === "error" || edit.binary) {
+      continue;
+    }
+    added += edit.added;
+    deleted += edit.deleted;
+  }
+  return {
+    reasoningSteps,
+    toolCalls,
+    fileCount: fileEdits.length,
+    added,
+    deleted,
+    hasEditingFiles,
+    hasFailedFiles: fileEdits.length > 0 && failedFileCount === fileEdits.length,
+    primaryFilePath,
+  };
 }
 
 interface AgentActivityClusterProps {
@@ -46,24 +104,56 @@ export function AgentActivityCluster({
   hasBodyBelow,
 }: AgentActivityClusterProps) {
   const { t } = useTranslation();
-  const reasoningSteps = messages.filter(isReasoningOnlyAssistant).length;
-  const toolCalls = countToolCalls(messages);
+  const fileEdits = useMemo(
+    () => summarizeFileEdits(collectFileEdits(messages), isTurnStreaming),
+    [messages, isTurnStreaming],
+  );
+  const {
+    reasoningSteps,
+    toolCalls,
+    fileCount,
+    added,
+    deleted,
+    hasEditingFiles,
+    hasFailedFiles,
+    primaryFilePath,
+  } = countActivity(messages, fileEdits);
 
   const [userToggledOuter, setUserToggledOuter] = useState(false);
   const [outerOpenLocal, setOuterOpenLocal] = useState(false);
+  const activityScrollRef = useRef<HTMLDivElement>(null);
+  const activityContentRef = useRef<HTMLDivElement>(null);
+  const autoFollowActivityRef = useRef(true);
+  const scrollFrameRef = useRef<number | null>(null);
   /** Collapsed by default during “Working…” and after the turn; user expands to inspect traces. */
   const outerExpanded = userToggledOuter ? outerOpenLocal : false;
 
-  const headerBusy = isTurnStreaming;
+  const hasLiveEditingFiles = isTurnStreaming && hasEditingFiles;
+  const headerBusy = fileCount > 0 ? hasEditingFiles : isTurnStreaming;
 
-  const summary =
-    isTurnStreaming
+  const fileActivitySummary = fileCount > 0
+    ? fileCount === 1 && primaryFilePath
+      ? t(fileActivitySummaryKey(hasLiveEditingFiles, hasFailedFiles), {
+          file: shortFileName(primaryFilePath),
+          defaultValue: `${fileActivityVerb(hasLiveEditingFiles, hasFailedFiles)} {{file}}`,
+        })
+      : t(fileActivityManySummaryKey(hasLiveEditingFiles, hasFailedFiles), {
+          count: fileCount,
+          defaultValue: `${fileActivityVerb(hasLiveEditingFiles, hasFailedFiles)} {{count}} files`,
+        })
+    : "";
+
+  const summary = fileCount > 0
+    ? fileActivitySummary
+    : isTurnStreaming
       ? reasoningSteps > 0
         ? t("message.agentActivityLiveSummary", {
             reasoning: reasoningSteps,
             tools: toolCalls,
             defaultValue: "Working… · {{reasoning}} steps · {{tools}} tool calls",
           })
+        : toolCalls === 0 && fileCount > 0
+          ? t("message.agentActivityLiveFilesOnly", { defaultValue: "Working…" })
         : t("message.agentActivityLiveToolsOnly", {
             tools: toolCalls,
             defaultValue: "Working… · {{tools}} tool calls",
@@ -74,16 +164,73 @@ export function AgentActivityCluster({
             tools: toolCalls,
             defaultValue: "{{reasoning}} steps · {{tools}} tool calls",
           })
+        : toolCalls === 0 && fileCount > 0
+          ? t("message.agentActivityFilesOnly", { defaultValue: "File changes" })
         : t("message.agentActivityToolsOnly", {
             tools: toolCalls,
             defaultValue: "{{tools}} tool calls",
           });
 
+  const cancelActivityScrollFrame = useCallback(() => {
+    if (scrollFrameRef.current !== null) {
+      window.cancelAnimationFrame(scrollFrameRef.current);
+      scrollFrameRef.current = null;
+    }
+  }, []);
+
+  const scrollActivityToBottom = useCallback(() => {
+    const el = activityScrollRef.current;
+    if (!el) return;
+    el.scrollTop = Math.max(0, el.scrollHeight - el.clientHeight);
+  }, []);
+
+  const scheduleActivityScrollToBottom = useCallback(() => {
+    cancelActivityScrollFrame();
+    scrollFrameRef.current = window.requestAnimationFrame(() => {
+      scrollFrameRef.current = null;
+      scrollActivityToBottom();
+    });
+  }, [cancelActivityScrollFrame, scrollActivityToBottom]);
+
   const toggleOuter = () => {
+    const nextOpen = userToggledOuter ? !outerOpenLocal : !outerExpanded;
+    if (nextOpen) {
+      autoFollowActivityRef.current = true;
+    }
     setUserToggledOuter(true);
-    setOuterOpenLocal((v) => (userToggledOuter ? !v : !outerExpanded));
+    setOuterOpenLocal(nextOpen);
   };
 
+  useLayoutEffect(() => {
+    if (!outerExpanded || !autoFollowActivityRef.current) return;
+    scheduleActivityScrollToBottom();
+  }, [outerExpanded, messages, isTurnStreaming, scheduleActivityScrollToBottom]);
+
+  useEffect(() => {
+    if (!outerExpanded) {
+      autoFollowActivityRef.current = true;
+      return;
+    }
+    const target = activityContentRef.current;
+    if (!target || typeof ResizeObserver === "undefined") return;
+    const observer = new ResizeObserver(() => {
+      if (autoFollowActivityRef.current) {
+        scheduleActivityScrollToBottom();
+      }
+    });
+    observer.observe(target);
+    return () => observer.disconnect();
+  }, [outerExpanded, scheduleActivityScrollToBottom]);
+
+  useEffect(() => cancelActivityScrollFrame, [cancelActivityScrollFrame]);
+
+  const onActivityScroll = useCallback(() => {
+    const el = activityScrollRef.current;
+    if (!el) return;
+    const distance = el.scrollHeight - el.scrollTop - el.clientHeight;
+    autoFollowActivityRef.current = distance < ACTIVITY_SCROLL_NEAR_BOTTOM_PX;
+  }, []);
+
   return (
     <div className={cn("w-full", hasBodyBelow && "mb-2")}>
       <button
@@ -96,12 +243,19 @@ export function AgentActivityCluster({
         aria-expanded={outerExpanded}
       >
         <Layers className="h-3.5 w-3.5 shrink-0" aria-hidden />
-        <StreamingLabelSheen
-          active={headerBusy}
-          className="min-w-0 flex-1 text-left"
-        >
-          {summary}
-        </StreamingLabelSheen>
+        <span className="flex min-w-0 flex-1 flex-wrap items-center gap-x-1.5 gap-y-0.5 text-left">
+          <StreamingLabelSheen
+            active={headerBusy}
+            className="min-w-0"
+          >
+            {summary}
+          </StreamingLabelSheen>
+          {fileCount > 0 && (
+            <span className="inline-flex min-w-0 items-center gap-1 text-muted-foreground/85">
+              <DiffPair added={added} deleted={deleted} />
+            </span>
+          )}
+        </span>
         <ChevronRight
           aria-hidden
           className={cn(
@@ -118,29 +272,38 @@ export function AgentActivityCluster({
           )}
         >
           <div
+            ref={activityScrollRef}
+            data-testid="agent-activity-scroll"
+            onScroll={onActivityScroll}
             className={cn(
               CLUSTER_SCROLL_MAX_CLASS,
               "overflow-y-auto px-2 py-1.5 scrollbar-thin scrollbar-track-transparent",
             )}
           >
-            <div className="flex flex-col gap-2">
+            <div ref={activityContentRef} className="flex flex-col gap-2">
               {messages.map((m) => {
                 if (isReasoningOnlyAssistant(m)) {
                   return (
                     <ReasoningBubble
                       key={m.id}
                       text={m.reasoning ?? ""}
-                      streaming={!!m.reasoningStreaming}
+                      streaming={isTurnStreaming && !!m.reasoningStreaming}
                       hasBodyBelow={false}
                       embeddedInCluster
                     />
                   );
                 }
                 if (m.kind === "trace") {
-                  return <TraceGroup key={m.id} message={m} animClass="" />;
+                  const hasTraceLines = (m.traces?.length ?? 0) > 0 || m.content.trim().length > 0;
+                  return hasTraceLines ? (
+                    <div key={m.id} className="flex flex-col gap-1">
+                      <TraceGroup message={m} animClass="" />
+                    </div>
+                  ) : null;
                 }
                 return null;
               })}
+              {fileEdits.length ? <FileEditGroup edits={fileEdits} /> : null}
             </div>
           </div>
         </div>
@@ -148,3 +311,231 @@ export function AgentActivityCluster({
     </div>
   );
 }
+
+function shortFileName(path: string): string {
+  return path.split(/[\\/]/).pop() || path;
+}
+
+function fileActivityVerb(editing: boolean, failed: boolean): string {
+  if (failed) return "Failed";
+  return editing ? "Editing" : "Edited";
+}
+
+function fileActivitySummaryKey(editing: boolean, failed: boolean): string {
+  if (failed) return "message.fileActivityFailedOne";
+  return editing ? "message.fileActivityEditingOne" : "message.fileActivityEditedOne";
+}
+
+function fileActivityManySummaryKey(editing: boolean, failed: boolean): string {
+  if (failed) return "message.fileActivityFailedMany";
+  return editing ? "message.fileActivityEditingMany" : "message.fileActivityEditedMany";
+}
+
+function fileEditCallKey(edit: UIFileEdit): string {
+  return `${edit.call_id}|${edit.tool}|${edit.path}`;
+}
+
+function collectFileEdits(messages: UIMessage[]): UIFileEdit[] {
+  const edits: UIFileEdit[] = [];
+  for (const message of messages) {
+    if (message.kind === "trace" && message.fileEdits?.length) {
+      edits.push(...message.fileEdits);
+    }
+  }
+  return edits;
+}
+
+function latestFileEditEvents(edits: UIFileEdit[]): UIFileEdit[] {
+  const order: string[] = [];
+  const byKey = new Map<string, UIFileEdit>();
+  for (const edit of edits) {
+    const key = fileEditCallKey(edit);
+    if (!byKey.has(key)) order.push(key);
+    byKey.set(key, edit);
+  }
+  return order.map((key) => byKey.get(key)).filter(Boolean) as UIFileEdit[];
+}
+
+function summarizeFileEdits(edits: UIFileEdit[], active: boolean): FileEditSummary[] {
+  interface MutableSummary {
+    key: string;
+    path: string;
+    added: number;
+    deleted: number;
+    approximate: boolean;
+    binary: boolean;
+    hasSuccessfulChange: boolean;
+    hasActiveEditing: boolean;
+    hasFailed: boolean;
+    error?: string;
+  }
+
+  const order: string[] = [];
+  const byPath = new Map<string, MutableSummary>();
+  for (const edit of latestFileEditEvents(edits)) {
+    const key = edit.path;
+    let summary = byPath.get(key);
+    if (!summary) {
+      summary = {
+        key,
+        path: edit.path,
+        added: 0,
+        deleted: 0,
+        approximate: false,
+        binary: false,
+        hasSuccessfulChange: false,
+        hasActiveEditing: false,
+        hasFailed: false,
+      };
+      byPath.set(key, summary);
+      order.push(key);
+    }
+
+    if (active && edit.status === "editing") {
+      summary.hasActiveEditing = true;
+      summary.binary = summary.binary || !!edit.binary;
+      summary.approximate = summary.approximate || !!edit.approximate;
+      if (!edit.binary) {
+        summary.added += edit.added;
+        summary.deleted += edit.deleted;
+      }
+      continue;
+    }
+
+    if (edit.status === "error") {
+      summary.hasFailed = true;
+      summary.error = edit.error ?? summary.error;
+      continue;
+    }
+
+    summary.hasSuccessfulChange = true;
+    summary.binary = summary.binary || !!edit.binary;
+    summary.approximate = active && (summary.approximate || !!edit.approximate);
+    if (!edit.binary) {
+      summary.added += edit.added;
+      summary.deleted += edit.deleted;
+    }
+  }
+
+  return order.map((key) => {
+    const summary = byPath.get(key)!;
+    const status: UIFileEdit["status"] = summary.hasActiveEditing
+      ? "editing"
+      : summary.hasSuccessfulChange
+        ? "done"
+        : summary.hasFailed
+          ? "error"
+          : "done";
+    return {
+      key: summary.key,
+      path: summary.path,
+      added: summary.added,
+      deleted: summary.deleted,
+      approximate: summary.approximate,
+      binary: summary.binary,
+      status,
+      error: summary.error,
+    };
+  });
+}
+
+function FileEditGroup({ edits }: { edits: FileEditSummary[] }) {
+  if (edits.length === 0) return null;
+  return (
+    <ul className="space-y-1 border-l border-muted-foreground/15 pl-3">
+      {edits.map((edit) => (
+        <FileEditRow key={edit.key} edit={edit} />
+      ))}
+    </ul>
+  );
+}
+
+function FileEditRow({ edit }: { edit: FileEditSummary }) {
+  const { t } = useTranslation();
+  const editing = edit.status === "editing";
+  const failed = edit.status === "error";
+  const hasCountedDiff = !failed && !edit.binary;
+  return (
+    <li className="grid grid-cols-[minmax(0,1fr)_auto] items-center gap-3 rounded-md px-2 py-1.5 text-xs">
+      <div className="flex min-w-0 items-center gap-2">
+        <FileReferenceChip
+          path={edit.path}
+          display="path"
+          active={editing}
+          className="min-w-0"
+          textClassName="text-[12px]"
+          testId="activity-file-reference"
+        />
+        {failed ? (
+          <span className="inline-flex shrink-0 items-center gap-1 text-[10.5px] font-medium text-destructive/75">
+            <AlertCircle className="h-3 w-3" aria-hidden />
+            {t("message.fileEditFailed", { defaultValue: "Failed" })}
+          </span>
+        ) : null}
+        {edit.approximate && !failed ? (
+          <span className="shrink-0 text-[10.5px] font-medium text-muted-foreground/55">
+            {t("message.fileEditApproximate", { defaultValue: "estimated" })}
+          </span>
+        ) : null}
+      </div>
+      {hasCountedDiff ? (
+        <DiffPair added={edit.added} deleted={edit.deleted} />
+      ) : null}
+    </li>
+  );
+}
+
+function DiffPair({ added, deleted }: { added: number; deleted: number }) {
+  return (
+    <span className="inline-flex shrink-0 items-center gap-1.5 tabular-nums">
+      <span className="text-emerald-600/75 dark:text-emerald-300/75">
+        +<AnimatedNumber value={added} />
+      </span>
+      <span className="text-rose-600/70 dark:text-rose-300/75">
+        -<AnimatedNumber value={deleted} />
+      </span>
+    </span>
+  );
+}
+
+function AnimatedNumber({ value }: { value: number }) {
+  const safeValue = Number.isFinite(value) ? Math.max(0, Math.round(value)) : 0;
+  const [display, setDisplay] = useState(0);
+  const displayRef = useRef(0);
+
+  const setAnimatedDisplay = useCallback((next: number) => {
+    displayRef.current = next;
+    setDisplay(next);
+  }, []);
+
+  useEffect(() => {
+    const reduceMotion = window.matchMedia?.("(prefers-reduced-motion: reduce)").matches;
+    if (reduceMotion) {
+      setAnimatedDisplay(safeValue);
+      return;
+    }
+    const start = displayRef.current;
+    const delta = safeValue - start;
+    if (delta === 0) {
+      setAnimatedDisplay(safeValue);
+      return;
+    }
+    const duration = 260;
+    const startedAt = performance.now();
+    let frame = 0;
+    const tick = (now: number) => {
+      const progress = Math.min(1, (now - startedAt) / duration);
+      const eased = 1 - Math.pow(1 - progress, 3);
+      setAnimatedDisplay(Math.round(start + delta * eased));
+      if (progress < 1) {
+        frame = window.requestAnimationFrame(tick);
+        return;
+      }
+      displayRef.current = safeValue;
+    };
+    frame = window.requestAnimationFrame(tick);
+    return () => window.cancelAnimationFrame(frame);
+  }, [safeValue, setAnimatedDisplay]);
+
+  return <>{display}</>;
+}

webui/src/components/thread/ThreadMessages.tsx

@@ -1,3 +1,6 @@
+import { useMemo } from "react";
+import { useTranslation } from "react-i18next";
+
 import { MessageBubble } from "@/components/MessageBubble";
 import {
   AgentActivityCluster,
@@ -9,6 +12,8 @@ interface ThreadMessagesProps {
   messages: UIMessage[];
   /** When true, agent turn still in flight — keeps activity cluster expanded. */
   isStreaming?: boolean;
+  hiddenMessageCount?: number;
+  onLoadEarlier?: () => void;
 }
 
 export type DisplayUnit =
@@ -30,31 +35,160 @@ export function isFinalAssistantSliceBeforeNextUser(
   return true;
 }
 
-function buildDisplayUnits(messages: UIMessage[]): DisplayUnit[] {
+export function buildDisplayUnits(messages: UIMessage[]): DisplayUnit[] {
   const out: DisplayUnit[] = [];
   let i = 0;
   while (i < messages.length) {
     const m = messages[i];
     if (isAgentActivityMember(m)) {
       const cluster: UIMessage[] = [];
-      while (i < messages.length && isAgentActivityMember(messages[i])) {
-        cluster.push(messages[i]);
+      let segmentId: string | undefined = m.activitySegmentId;
+      let clusterHasFileEdits = hasFileEdits(m);
+      while (
+        i < messages.length
+        && isAgentActivityMember(messages[i])
+        && canJoinActivityCluster(segmentId, clusterHasFileEdits, messages[i])
+      ) {
+        const current = messages[i];
+        if (!segmentId && current.activitySegmentId) {
+          segmentId = current.activitySegmentId;
+        }
+        clusterHasFileEdits = clusterHasFileEdits || hasFileEdits(current);
+        cluster.push(current);
         i += 1;
       }
       out.push({ type: "cluster", messages: cluster });
       continue;
     }
+    const previous = out[out.length - 1];
+    if (
+      previous?.type === "cluster"
+      && assistantHasInlineReasoning(m)
+      && canFoldInlineReasoning(previous.messages, m)
+    ) {
+      previous.messages.push(reasoningOnlyMessageFromAnswer(m));
+      out.push({ type: "single", message: stripInlineReasoning(m) });
+      i += 1;
+      continue;
+    }
+    if (assistantHasInlineReasoning(m)) {
+      out.push({ type: "cluster", messages: [reasoningOnlyMessageFromAnswer(m)] });
+      out.push({ type: "single", message: stripInlineReasoning(m) });
+      i += 1;
+      continue;
+    }
     out.push({ type: "single", message: m });
     i += 1;
   }
   return out;
 }
 
-export function ThreadMessages({ messages, isStreaming = false }: ThreadMessagesProps) {
-  const units = buildDisplayUnits(messages);
+function clusterSegmentId(messages: UIMessage[]): string | undefined {
+  return messages.find((message) => message.activitySegmentId)?.activitySegmentId;
+}
+
+function hasFileEdits(message: UIMessage): boolean {
+  return !!message.fileEdits?.length;
+}
+
+function clusterHasFileEdits(messages: UIMessage[]): boolean {
+  return messages.some(hasFileEdits);
+}
+
+function canJoinActivityCluster(
+  clusterSegmentId: string | undefined,
+  clusterIncludesFileEdits: boolean,
+  message: UIMessage,
+): boolean {
+  const messageHasFileEdits = hasFileEdits(message);
+  if (!clusterIncludesFileEdits && !messageHasFileEdits) return true;
+  if (!clusterSegmentId || !message.activitySegmentId) return true;
+  return clusterSegmentId === message.activitySegmentId;
+}
+
+function canFoldInlineReasoning(cluster: UIMessage[], message: UIMessage): boolean {
+  if (!clusterHasFileEdits(cluster) && !hasFileEdits(message)) return true;
+  const segmentId = clusterSegmentId(cluster);
+  if (!segmentId || !message.activitySegmentId) return true;
+  return segmentId === message.activitySegmentId;
+}
+
+function assistantHasInlineReasoning(message: UIMessage): boolean {
+  return (
+    message.role === "assistant"
+    && message.kind !== "trace"
+    && message.content.trim().length > 0
+    && (!!message.reasoning?.trim() || !!message.reasoningStreaming)
+  );
+}
+
+function reasoningOnlyMessageFromAnswer(message: UIMessage): UIMessage {
+  return {
+    id: `${message.id}-reasoning`,
+    role: "assistant",
+    content: "",
+    createdAt: message.createdAt,
+    reasoning: message.reasoning,
+    reasoningStreaming: message.reasoningStreaming,
+    isStreaming: message.reasoningStreaming,
+    activitySegmentId: message.activitySegmentId,
+  };
+}
+
+function stripInlineReasoning(message: UIMessage): UIMessage {
+  const next = { ...message };
+  delete next.reasoning;
+  delete next.reasoningStreaming;
+  return next;
+}
+
+export function assistantCopyFlags(units: DisplayUnit[]): boolean[] {
+  const flags = new Array<boolean>(units.length).fill(true);
+  let hasLaterUnitBeforeUser = false;
+  for (let i = units.length - 1; i >= 0; i -= 1) {
+    const unit = units[i];
+    if (unit.type === "single" && unit.message.role === "user") {
+      hasLaterUnitBeforeUser = false;
+      continue;
+    }
+    if (unit.type === "single" && unit.message.role === "assistant") {
+      flags[i] = !hasLaterUnitBeforeUser;
+    }
+    hasLaterUnitBeforeUser = true;
+  }
+  return flags;
+}
+
+export function ThreadMessages({
+  messages,
+  isStreaming = false,
+  hiddenMessageCount = 0,
+  onLoadEarlier,
+}: ThreadMessagesProps) {
+  const { t } = useTranslation();
+  const units = useMemo(() => buildDisplayUnits(messages), [messages]);
+  const copyFlags = useMemo(() => assistantCopyFlags(units), [units]);
+  const liveActivityClusterIndex = useMemo(
+    () => isStreaming ? currentActivityClusterIndex(units) : -1,
+    [isStreaming, units],
+  );
 
   return (
     <div className="flex w-full flex-col">
+      {hiddenMessageCount > 0 && onLoadEarlier ? (
+        <div className="mb-4 flex justify-center">
+          <button
+            type="button"
+            onClick={onLoadEarlier}
+            className="rounded-full border border-border/60 bg-background/85 px-3 py-1.5 text-xs font-medium text-muted-foreground shadow-sm transition-colors hover:bg-muted/55 hover:text-foreground"
+          >
+            {t("thread.loadEarlier", {
+              count: hiddenMessageCount,
+              defaultValue: "Load earlier messages",
+            })}
+          </button>
+        </div>
+      ) : null}
       {units.map((unit, index) => {
         const prev = units[index - 1];
         const marginTop =
@@ -72,7 +206,7 @@ export function ThreadMessages({ messages, isStreaming = false }: ThreadMessages
             {unit.type === "cluster" ? (
               <AgentActivityCluster
                 messages={unit.messages}
-                isTurnStreaming={isStreaming}
+                isTurnStreaming={index === liveActivityClusterIndex}
                 hasBodyBelow={hasBodyBelow}
               />
             ) : (
@@ -80,7 +214,7 @@ export function ThreadMessages({ messages, isStreaming = false }: ThreadMessages
                 message={unit.message}
                 showAssistantCopyAction={
                   unit.message.role === "assistant"
-                    ? isFinalAssistantSliceBeforeNextUser(units, index)
+                    ? copyFlags[index]
                     : true
                 }
               />
@@ -92,6 +226,11 @@ export function ThreadMessages({ messages, isStreaming = false }: ThreadMessages
   );
 }
 
+function currentActivityClusterIndex(units: DisplayUnit[]): number {
+  const last = units.length - 1;
+  return units[last]?.type === "cluster" ? last : -1;
+}
+
 function unitKey(unit: DisplayUnit, index: number): string {
   if (unit.type === "cluster") {
     const anchor = unit.messages[0]?.id;

webui/src/components/thread/ThreadShell.tsx

@@ -167,8 +167,9 @@ export function ThreadShell({
 
   useEffect(() => {
     if (!chatId) return;
-    return client.onSessionUpdate((updatedChatId) => {
+    return client.onSessionUpdate((updatedChatId, scope) => {
       if (updatedChatId !== chatId) return;
+      if (scope === "metadata") return;
       pendingCanonicalHydrateRef.current.add(chatId);
       refreshHistory();
     });
@@ -389,6 +390,7 @@ export function ThreadShell({
         composer={composer}
         scrollToBottomSignal={scrollToBottomSignal}
         conversationKey={historyKey}
+        showScrollToBottomButton={!!session}
       />
     </section>
   );

webui/src/components/thread/ThreadViewport.tsx

@@ -1,8 +1,17 @@
-import { type ReactNode, useCallback, useEffect, useLayoutEffect, useRef, useState } from "react";
+import {
+  type ReactNode,
+  useCallback,
+  useEffect,
+  useLayoutEffect,
+  useMemo,
+  useRef,
+  useState,
+} from "react";
 import { ArrowDown } from "lucide-react";
 import { useTranslation } from "react-i18next";
 
 import { ThreadMessages } from "@/components/thread/ThreadMessages";
+import { isAgentActivityMember } from "@/components/thread/AgentActivityCluster";
 import { Button } from "@/components/ui/button";
 import { cn } from "@/lib/utils";
 import type { UIMessage } from "@/lib/types";
@@ -14,9 +23,27 @@ interface ThreadViewportProps {
   emptyState?: ReactNode;
   scrollToBottomSignal?: number;
   conversationKey?: string | null;
+  showScrollToBottomButton?: boolean;
 }
 
 const NEAR_BOTTOM_PX = 48;
+const DEFAULT_SCROLL_BUTTON_BOTTOM_PX = 192;
+const SCROLL_BUTTON_COMPOSER_GAP_PX = 16;
+export const INITIAL_HISTORY_WINDOW = 160;
+export const HISTORY_WINDOW_INCREMENT = 120;
+
+export function windowMessages(messages: UIMessage[], visibleCount: number): UIMessage[] {
+  if (messages.length <= visibleCount) return messages;
+  let start = Math.max(0, messages.length - visibleCount);
+  while (
+    start > 0
+    && isAgentActivityMember(messages[start])
+    && isAgentActivityMember(messages[start - 1])
+  ) {
+    start -= 1;
+  }
+  return messages.slice(start);
+}
 
 export function ThreadViewport({
   messages,
@@ -25,18 +52,33 @@ export function ThreadViewport({
   emptyState,
   scrollToBottomSignal = 0,
   conversationKey = null,
+  showScrollToBottomButton = true,
 }: ThreadViewportProps) {
   const { t } = useTranslation();
   const scrollRef = useRef<HTMLDivElement>(null);
   const contentRef = useRef<HTMLDivElement>(null);
+  const composerDockRef = useRef<HTMLDivElement>(null);
   const bottomRef = useRef<HTMLDivElement>(null);
   const lastConversationKeyRef = useRef<string | null>(conversationKey);
   const pendingConversationScrollRef = useRef(true);
   const scrollFrameIdsRef = useRef<number[]>([]);
+  const restoreScrollAfterPrependRef =
+    useRef<{ height: number; top: number } | null>(null);
   /** User scrolled away from the bottom; do not auto-yank until they return or we reset (new chat / send). */
   const userReadingHistoryRef = useRef(false);
   const [atBottom, setAtBottom] = useState(true);
+  const [composerDockHeight, setComposerDockHeight] = useState(0);
+  const [visibleMessageCount, setVisibleMessageCount] =
+    useState(INITIAL_HISTORY_WINDOW);
   const hasMessages = messages.length > 0;
+  const visibleMessages = useMemo(
+    () => windowMessages(messages, visibleMessageCount),
+    [messages, visibleMessageCount],
+  );
+  const hiddenMessageCount = messages.length - visibleMessages.length;
+  const scrollButtonBottom = composerDockHeight > 0
+    ? composerDockHeight + SCROLL_BUTTON_COMPOSER_GAP_PX
+    : DEFAULT_SCROLL_BUTTON_BOTTOM_PX;
 
   const cancelScheduledBottomScroll = useCallback(() => {
     for (const id of scrollFrameIdsRef.current) {
@@ -77,6 +119,30 @@ export function ThreadViewport({
     [cancelScheduledBottomScroll, scrollToBottomNow],
   );
 
+  const loadEarlierMessages = useCallback(() => {
+    const el = scrollRef.current;
+    if (el) {
+      restoreScrollAfterPrependRef.current = {
+        height: el.scrollHeight,
+        top: el.scrollTop,
+      };
+    }
+    userReadingHistoryRef.current = true;
+    setAtBottom(false);
+    setVisibleMessageCount((count) =>
+      Math.min(messages.length, count + HISTORY_WINDOW_INCREMENT),
+    );
+  }, [messages.length]);
+
+  const measureComposerDock = useCallback(() => {
+    const el = composerDockRef.current;
+    if (!el) return;
+    const height = el.getBoundingClientRect().height || el.offsetHeight;
+    setComposerDockHeight((current) =>
+      Math.abs(current - height) < 1 ? current : height,
+    );
+  }, []);
+
   useEffect(() => {
     if (!atBottom) return;
     // Instant jump: CSS scroll-smooth + behavior "auto" still animates in some
@@ -96,8 +162,19 @@ export function ThreadViewport({
     pendingConversationScrollRef.current = true;
     userReadingHistoryRef.current = false;
     setAtBottom(true);
+    setVisibleMessageCount(INITIAL_HISTORY_WINDOW);
   }, [conversationKey]);
 
+  useLayoutEffect(() => {
+    const pending = restoreScrollAfterPrependRef.current;
+    if (!pending) return;
+    const el = scrollRef.current;
+    restoreScrollAfterPrependRef.current = null;
+    if (!el) return;
+    const delta = el.scrollHeight - pending.height;
+    el.scrollTop = pending.top + delta;
+  }, [visibleMessages.length]);
+
   useLayoutEffect(() => {
     if (!pendingConversationScrollRef.current) return;
     if (!conversationKey) {
@@ -110,6 +187,10 @@ export function ThreadViewport({
     pendingConversationScrollRef.current = false;
   }, [conversationKey, hasMessages, messages, scrollToBottom]);
 
+  useLayoutEffect(() => {
+    measureComposerDock();
+  }, [composer, hasMessages, measureComposerDock]);
+
   useEffect(() => cancelScheduledBottomScroll, [cancelScheduledBottomScroll]);
 
   useEffect(() => {
@@ -123,6 +204,14 @@ export function ThreadViewport({
     return () => observer.disconnect();
   }, [hasMessages, scrollToBottom]);
 
+  useEffect(() => {
+    const target = composerDockRef.current;
+    if (!target || typeof ResizeObserver === "undefined") return;
+    const observer = new ResizeObserver(() => measureComposerDock());
+    observer.observe(target);
+    return () => observer.disconnect();
+  }, [hasMessages, measureComposerDock]);
+
   useEffect(() => {
     const el = scrollRef.current;
     if (!el) return;
@@ -155,11 +244,20 @@ export function ThreadViewport({
           <div ref={contentRef} className="mx-auto flex min-h-full w-full max-w-[64rem] flex-col">
             <div className="flex-1 px-4 pb-20 pt-4">
               <div className="mx-auto w-full max-w-[49.5rem]">
-                <ThreadMessages messages={messages} isStreaming={isStreaming} />
+                <ThreadMessages
+                  messages={visibleMessages}
+                  isStreaming={isStreaming}
+                  hiddenMessageCount={hiddenMessageCount}
+                  onLoadEarlier={loadEarlierMessages}
+                />
               </div>
             </div>
 
-            <div className="sticky bottom-0 z-10 mt-auto bg-background">
+            <div
+              ref={composerDockRef}
+              data-testid="thread-composer-dock"
+              className="sticky bottom-0 z-10 mt-auto bg-background"
+            >
               <div className="px-4 pb-3">
                 {composer}
               </div>
@@ -183,17 +281,18 @@ export function ThreadViewport({
         className="pointer-events-none absolute inset-x-0 top-0 h-6 bg-gradient-to-b from-background to-transparent"
       />
 
-      {!atBottom && (
+      {showScrollToBottomButton && !atBottom && (
         <Button
           variant="outline"
           size="icon"
           onClick={() => scrollToBottom(true, 1, { force: true })}
           className={cn(
             /* Keep clear of sticky composer (textarea + toolbar + optional goal strip). */
-            "absolute bottom-48 left-1/2 z-20 h-8 w-8 -translate-x-1/2 rounded-full shadow-md",
+            "absolute left-1/2 z-20 h-8 w-8 -translate-x-1/2 rounded-full shadow-md",
             "bg-background/90 backdrop-blur",
             "animate-in fade-in-0 zoom-in-95",
           )}
+          style={{ bottom: scrollButtonBottom }}
           aria-label={t("thread.scrollToBottom")}
         >
           <ArrowDown className="h-4 w-4" />

webui/src/components/ui/tooltip.tsx

@@ -11,15 +11,17 @@ const TooltipContent = React.forwardRef<
   React.ElementRef<typeof TooltipPrimitive.Content>,
   React.ComponentPropsWithoutRef<typeof TooltipPrimitive.Content>
 >(({ className, sideOffset = 4, ...props }, ref) => (
-  <TooltipPrimitive.Content
-    ref={ref}
-    sideOffset={sideOffset}
-    className={cn(
-      "z-50 overflow-hidden rounded-md border bg-popover px-3 py-1.5 text-xs text-popover-foreground shadow-md animate-in fade-in-0 zoom-in-95",
-      className,
-    )}
-    {...props}
-  />
+  <TooltipPrimitive.Portal>
+    <TooltipPrimitive.Content
+      ref={ref}
+      sideOffset={sideOffset}
+      className={cn(
+        "z-50 overflow-hidden rounded-md border bg-popover px-3 py-1.5 text-xs text-popover-foreground shadow-md animate-in fade-in-0 zoom-in-95",
+        className,
+      )}
+      {...props}
+    />
+  </TooltipPrimitive.Portal>
 ));
 TooltipContent.displayName = TooltipPrimitive.Content.displayName;
 

webui/src/globals.css

@@ -117,53 +117,60 @@
     --cjk-line-height: 1.625;
   }
 
-  /* L→R sheen over solid label text (overlay stripe). Avoids ``background-clip:
-     text`` loop seams that read as RTL “erase” or one-frame transparent glyphs. */
-  @keyframes reasoning-sheen-ltr {
+  /* L→R sheen clipped to live activity labels. The highlight lives inside
+     the glyphs, not in the row background, so dark mode stays quiet. */
+  @keyframes streaming-text-sheen-ltr {
     0% {
-      left: -44%;
+      background-position: 140% 50%;
     }
     100% {
-      left: 118%;
+      background-position: -40% 50%;
     }
   }
-  .reasoning-sheen-track {
+  .streaming-text-sheen {
+    position: relative;
+    color: hsl(var(--muted-foreground));
+  }
+  .streaming-text-sheen::after {
+    content: attr(data-sheen-text);
     position: absolute;
     inset: 0;
-    z-index: 1;
+    display: block;
     overflow: hidden;
-    border-radius: 2px;
+    white-space: nowrap;
+    text-overflow: ellipsis;
     pointer-events: none;
-  }
-  .reasoning-sheen-stripe {
-    position: absolute;
-    top: 0;
-    bottom: 0;
-    width: 44%;
-    min-width: 3.25rem;
-    left: -44%;
-    border-radius: inherit;
+    color: transparent;
     background: linear-gradient(
       90deg,
       transparent 0%,
-      hsl(0 0% 100% / 0.07) 34%,
-      hsl(0 0% 100% / 0.76) 50%,
-      hsl(0 0% 100% / 0.07) 66%,
+      transparent 38%,
+      hsl(var(--foreground) / 0.98) 50%,
+      transparent 62%,
       transparent 100%
     );
-    mix-blend-mode: soft-light;
-    opacity: 0.95;
-    animation: reasoning-sheen-ltr 5.2s linear infinite;
+    background-size: 260% 100%;
+    background-position: 140% 50%;
+    background-repeat: no-repeat;
+    background-clip: text;
+    -webkit-background-clip: text;
+    -webkit-text-fill-color: transparent;
+    animation: streaming-text-sheen-ltr 2.8s ease-in-out infinite;
   }
-  .dark .reasoning-sheen-stripe {
-    mix-blend-mode: overlay;
-    opacity: 1;
+  .dark .streaming-text-sheen::after {
+    background-image: linear-gradient(
+      90deg,
+      transparent 0%,
+      transparent 38%,
+      hsl(var(--foreground) / 0.98) 50%,
+      transparent 62%,
+      transparent 100%
+    );
   }
   @media (prefers-reduced-motion: reduce) {
-    .reasoning-sheen-stripe {
+    .streaming-text-sheen::after {
       animation: none;
-      opacity: 0;
-      visibility: hidden;
+      content: "";
     }
   }
 

webui/src/hooks/useDeferredTitleRefresh.ts

@@ -0,0 +1,68 @@
+import { useCallback, useEffect, useRef } from "react";
+
+import type { ChatSummary } from "@/lib/types";
+
+const TITLE_REFRESH_RETRY_DELAYS_MS = [1_000, 3_000, 7_000] as const;
+
+function hasGeneratedTitle(session: ChatSummary | null): boolean {
+  return !!session?.title?.trim();
+}
+
+/**
+ * The server generates WebUI titles after the main turn has already ended.
+ * Refresh once immediately, then retry lightly for untitled sessions so the
+ * async title appears even if the websocket metadata notification is delayed.
+ */
+export function useDeferredTitleRefresh(
+  activeSession: ChatSummary | null,
+  refresh: () => Promise<void>,
+  retryDelaysMs: readonly number[] = TITLE_REFRESH_RETRY_DELAYS_MS,
+): () => void {
+  const activeSessionRef = useRef(activeSession);
+  const timersRef = useRef<ReturnType<typeof setTimeout>[]>([]);
+  activeSessionRef.current = activeSession;
+
+  const clearTimers = useCallback(() => {
+    for (const timer of timersRef.current) {
+      clearTimeout(timer);
+    }
+    timersRef.current = [];
+  }, []);
+
+  useEffect(() => clearTimers, [clearTimers]);
+
+  useEffect(() => {
+    clearTimers();
+  }, [activeSession?.key, clearTimers]);
+
+  useEffect(() => {
+    if (hasGeneratedTitle(activeSession)) {
+      clearTimers();
+    }
+  }, [activeSession, clearTimers]);
+
+  return useCallback(() => {
+    void refresh();
+
+    const sessionAtTurnEnd = activeSessionRef.current;
+    if (!sessionAtTurnEnd || hasGeneratedTitle(sessionAtTurnEnd)) {
+      return;
+    }
+
+    clearTimers();
+    for (const delayMs of retryDelaysMs) {
+      const timer = setTimeout(() => {
+        const latest = activeSessionRef.current;
+        if (
+          !latest ||
+          latest.key !== sessionAtTurnEnd.key ||
+          hasGeneratedTitle(latest)
+        ) {
+          return;
+        }
+        void refresh();
+      }, delayMs);
+      timersRef.current.push(timer);
+    }
+  }, [clearTimers, refresh, retryDelaysMs]);
+}

webui/src/hooks/useNanobotStream.ts

@@ -10,6 +10,7 @@ import type {
   OutboundMedia,
   GoalStateWsPayload,
   UIImage,
+  UIFileEdit,
   UIMessage,
 } from "@/lib/types";
 
@@ -18,12 +19,26 @@ interface StreamBuffer {
   messageId: string;
 }
 
-/** Scan upward from the bottom skipping trace rows so tool breadcrumbs don't steal the stream target. */
-function findStreamingAssistantId(prev: UIMessage[]): string | null {
+interface ActiveAssistantCursor {
+  id: string;
+  index: number;
+}
+
+type PendingStreamEvent =
+  | { kind: "delta"; text: string }
+  | { kind: "reasoning"; text: string };
+
+/** Find a still-open streamed assistant turn. Closed stream segments stay visible
+ * as streaming until ``turn_end`` for visual continuity, but they must not
+ * receive later delta segments. */
+function findStreamingAssistantIndex(
+  prev: UIMessage[],
+  closedStreamIds: ReadonlySet<string>,
+): number | null {
   for (let i = prev.length - 1; i >= 0; i -= 1) {
     const m = prev[i];
     if (m.kind === "trace") continue;
-    if (m.role === "assistant" && m.isStreaming) return m.id;
+    if (m.role === "assistant" && m.isStreaming && !closedStreamIds.has(m.id)) return i;
     if (m.role === "user") break;
   }
   return null;
@@ -38,7 +53,13 @@ function findStreamingAssistantId(prev: UIMessage[]): string | null {
  * case the reasoning still belongs to the same assistant turn and must render
  * above the answer, not as a new row below it.
  */
-function attachReasoningChunk(prev: UIMessage[], chunk: string): UIMessage[] {
+function attachReasoningChunk(
+  prev: UIMessage[],
+  chunk: string,
+  segments?: {
+    ensure: () => string;
+  },
+): UIMessage[] {
   for (let i = prev.length - 1; i >= 0; i -= 1) {
     const candidate = prev[i];
     // A user turn is a hard boundary: reasoning after it belongs to the new
@@ -49,6 +70,7 @@ function attachReasoningChunk(prev: UIMessage[], chunk: string): UIMessage[] {
     // that produced those tool calls.
     if (candidate.kind === "trace") break;
     if (candidate.role !== "assistant") continue;
+    const activitySegmentId = candidate.activitySegmentId ?? segments?.ensure();
     const hasAnswer = candidate.content.length > 0;
     if (
       candidate.reasoningStreaming
@@ -60,6 +82,7 @@ function attachReasoningChunk(prev: UIMessage[], chunk: string): UIMessage[] {
         ...candidate,
         reasoning: (candidate.reasoning ?? "") + chunk,
         reasoningStreaming: true,
+        ...(activitySegmentId ? { activitySegmentId } : {}),
       };
       return [...prev.slice(0, i), merged, ...prev.slice(i + 1)];
     }
@@ -68,11 +91,13 @@ function attachReasoningChunk(prev: UIMessage[], chunk: string): UIMessage[] {
         ...candidate,
         reasoning: chunk,
         reasoningStreaming: true,
+        ...(activitySegmentId ? { activitySegmentId } : {}),
       };
       return [...prev.slice(0, i), merged, ...prev.slice(i + 1)];
     }
     break;
   }
+  const activitySegmentId = segments?.ensure();
   return [
     ...prev,
     {
@@ -82,6 +107,7 @@ function attachReasoningChunk(prev: UIMessage[], chunk: string): UIMessage[] {
       isStreaming: true,
       reasoning: chunk,
       reasoningStreaming: true,
+      ...(activitySegmentId ? { activitySegmentId } : {}),
       createdAt: Date.now(),
     },
   ];
@@ -95,13 +121,19 @@ function attachReasoningChunk(prev: UIMessage[], chunk: string): UIMessage[] {
  * the model already produced an answer in a previous turn, so the new
  * delta belongs in a fresh row.
  */
-function findActiveAssistantPlaceholder(prev: UIMessage[]): string | null {
+function findActiveAssistantPlaceholderIndex(prev: UIMessage[]): number | null {
   const last = prev[prev.length - 1];
   if (!last) return null;
   if (last.role !== "assistant" || last.kind === "trace") return null;
   if (last.content.length > 0) return null;
   if (!last.isStreaming) return null;
-  return last.id;
+  return prev.length - 1;
+}
+
+function replaceMessageAt(prev: UIMessage[], index: number, message: UIMessage): UIMessage[] {
+  const next = prev.slice();
+  next[index] = message;
+  return next;
 }
 
 /**
@@ -182,6 +214,47 @@ function absorbCompleteAssistantMessage(
   ];
 }
 
+function fileEditKey(edit: Pick<UIFileEdit, "call_id" | "tool" | "path">): string {
+  return `${edit.call_id}|${edit.tool}|${edit.path}`;
+}
+
+function normalizeFileEdit(edit: UIFileEdit): UIFileEdit | null {
+  if (!edit || !edit.path || !edit.tool) return null;
+  const inferredStatus =
+    edit.phase === "error"
+      ? "error"
+      : edit.phase === "end"
+        ? "done"
+        : "editing";
+  return {
+    ...edit,
+    call_id: edit.call_id || `${edit.tool}:${edit.path}`,
+    added: Number.isFinite(edit.added) ? Math.max(0, Math.round(edit.added)) : 0,
+    deleted: Number.isFinite(edit.deleted) ? Math.max(0, Math.round(edit.deleted)) : 0,
+    status: edit.status === "error" || edit.status === "done" || edit.status === "editing"
+      ? edit.status
+      : inferredStatus,
+  };
+}
+
+function mergeFileEdits(existing: UIFileEdit[] | undefined, incoming: UIFileEdit[]): UIFileEdit[] {
+  const next = [...(existing ?? [])];
+  const indexByKey = new Map(next.map((edit, index) => [fileEditKey(edit), index]));
+  for (const raw of incoming) {
+    const edit = normalizeFileEdit(raw);
+    if (!edit) continue;
+    const key = fileEditKey(edit);
+    const existingIndex = indexByKey.get(key);
+    if (existingIndex === undefined) {
+      indexByKey.set(key, next.length);
+      next.push(edit);
+      continue;
+    }
+    next[existingIndex] = { ...next[existingIndex], ...edit };
+  }
+  return next;
+}
+
 /**
  * Subscribe to a chat by ID. Returns the in-memory message list for the chat,
  * a streaming flag, and a ``send`` function. Initial history must be seeded
@@ -239,6 +312,13 @@ export function useNanobotStream(
   const [goalState, setGoalState] = useState<GoalStateWsPayload | undefined>(undefined);
   const [streamError, setStreamError] = useState<StreamError | null>(null);
   const buffer = useRef<StreamBuffer | null>(null);
+  const activeAssistantRef = useRef<ActiveAssistantCursor | null>(null);
+  const closedAssistantStreamIdsRef = useRef<Set<string>>(new Set());
+  const activitySegmentRef = useRef<string | null>(null);
+  const fileEditSegmentRef = useRef<string | null>(null);
+  const activitySegmentCounterRef = useRef(0);
+  const pendingStreamEventsRef = useRef<PendingStreamEvent[]>([]);
+  const streamFrameRef = useRef<number | null>(null);
   const suppressStreamUntilTurnEndRef = useRef(false);
   /** Timer that defers ``isStreaming = false`` after ``stream_end``.
    *
@@ -255,6 +335,159 @@ export function useNanobotStream(
 
   const dismissStreamError = useCallback(() => setStreamError(null), []);
 
+  const clearPendingStreamWork = useCallback(() => {
+    if (streamFrameRef.current !== null) {
+      window.cancelAnimationFrame(streamFrameRef.current);
+      streamFrameRef.current = null;
+    }
+    pendingStreamEventsRef.current = [];
+  }, []);
+
+  const createActivitySegmentId = useCallback((activate = true) => {
+    activitySegmentCounterRef.current += 1;
+    const id = `activity-${activitySegmentCounterRef.current}`;
+    if (activate) activitySegmentRef.current = id;
+    return id;
+  }, []);
+
+  const freshActivitySegmentId = useCallback(
+    () => createActivitySegmentId(true),
+    [createActivitySegmentId],
+  );
+
+  const detachedActivitySegmentId = useCallback(
+    () => createActivitySegmentId(false),
+    [createActivitySegmentId],
+  );
+
+  const ensureActivitySegmentId = useCallback(() => {
+    if (activitySegmentRef.current) return activitySegmentRef.current;
+    return freshActivitySegmentId();
+  }, [freshActivitySegmentId]);
+
+  const clearActivitySegment = useCallback(() => {
+    activitySegmentRef.current = null;
+    fileEditSegmentRef.current = null;
+  }, []);
+
+  const closeActiveAssistantStream = useCallback(() => {
+    const closedStreamId = buffer.current?.messageId ?? activeAssistantRef.current?.id;
+    if (closedStreamId) closedAssistantStreamIdsRef.current.add(closedStreamId);
+    buffer.current = null;
+    activeAssistantRef.current = null;
+  }, []);
+
+  const resolveActiveAssistantIndex = useCallback((prev: UIMessage[]): number | null => {
+    const cursor = activeAssistantRef.current;
+    if (!cursor) return null;
+    const indexed = prev[cursor.index];
+    if (indexed?.id === cursor.id && indexed.role === "assistant" && indexed.kind !== "trace") {
+      return cursor.index;
+    }
+    const idx = prev.findIndex((m) => m.id === cursor.id);
+    if (idx === -1) {
+      activeAssistantRef.current = null;
+      return null;
+    }
+    const found = prev[idx];
+    if (found.role !== "assistant" || found.kind === "trace") {
+      activeAssistantRef.current = null;
+      return null;
+    }
+    activeAssistantRef.current = { id: cursor.id, index: idx };
+    return idx;
+  }, []);
+
+  const appendAnswerChunk = useCallback(
+    (prev: UIMessage[], chunk: string): UIMessage[] => {
+      let next = prev;
+      let targetIndex = resolveActiveAssistantIndex(next);
+
+      if (targetIndex === null) {
+        targetIndex = findActiveAssistantPlaceholderIndex(next);
+      }
+      if (targetIndex === null) {
+        targetIndex = findStreamingAssistantIndex(next, closedAssistantStreamIdsRef.current);
+      }
+      if (targetIndex === null) {
+        const id = crypto.randomUUID();
+        next = [
+          ...next,
+          {
+            id,
+            role: "assistant",
+            content: "",
+            isStreaming: true,
+            createdAt: Date.now(),
+          },
+        ];
+        targetIndex = next.length - 1;
+      }
+
+      const target = next[targetIndex];
+      const merged: UIMessage = {
+        ...target,
+        content: target.content + chunk,
+        isStreaming: true,
+      };
+      closedAssistantStreamIdsRef.current.delete(merged.id);
+      activeAssistantRef.current = { id: merged.id, index: targetIndex };
+      buffer.current = { messageId: merged.id };
+      return replaceMessageAt(next, targetIndex, merged);
+    },
+    [resolveActiveAssistantIndex],
+  );
+
+  const applyPendingStreamEvents = useCallback(
+    (prev: UIMessage[], events: PendingStreamEvent[]): UIMessage[] => {
+      let next = prev;
+      for (let i = 0; i < events.length;) {
+        const kind = events[i].kind;
+        let text = "";
+        while (i < events.length && events[i].kind === kind) {
+          text += events[i].text;
+          i += 1;
+        }
+        next = kind === "delta"
+          ? appendAnswerChunk(next, text)
+          : attachReasoningChunk(next, text, {
+              ensure: ensureActivitySegmentId,
+            });
+      }
+      return next;
+    },
+    [appendAnswerChunk, ensureActivitySegmentId],
+  );
+
+  const flushPendingStreamEvents = useCallback((options?: { closeAnswerSegment?: boolean }) => {
+    if (streamFrameRef.current !== null) {
+      window.cancelAnimationFrame(streamFrameRef.current);
+      streamFrameRef.current = null;
+    }
+    const events = pendingStreamEventsRef.current;
+    if (events.length === 0) {
+      if (options?.closeAnswerSegment) closeActiveAssistantStream();
+      return;
+    }
+    pendingStreamEventsRef.current = [];
+    setMessages((prev) => {
+      const next = applyPendingStreamEvents(prev, events);
+      if (options?.closeAnswerSegment) closeActiveAssistantStream();
+      return next;
+    });
+  }, [applyPendingStreamEvents, closeActiveAssistantStream]);
+
+  const schedulePendingStreamFlush = useCallback(() => {
+    if (streamFrameRef.current !== null) return;
+    streamFrameRef.current = window.requestAnimationFrame(() => {
+      streamFrameRef.current = null;
+      const events = pendingStreamEventsRef.current;
+      if (events.length === 0) return;
+      pendingStreamEventsRef.current = [];
+      setMessages((prev) => applyPendingStreamEvents(prev, events));
+    });
+  }, [applyPendingStreamEvents]);
+
   // Reset local state when switching chats. Do not reset on every
   // ``initialMessages`` update: a brand-new chat can receive an empty/404
   // history response after the optimistic first message has already rendered.
@@ -269,13 +502,17 @@ export function useNanobotStream(
     setRunStartedAt(chatId ? client.getRunStartedAt(chatId) : null);
     setGoalState(chatId ? client.getGoalState(chatId) : undefined);
     buffer.current = null;
+    activeAssistantRef.current = null;
+    closedAssistantStreamIdsRef.current.clear();
+    clearActivitySegment();
+    clearPendingStreamWork();
     suppressStreamUntilTurnEndRef.current = false;
     if (streamEndTimerRef.current !== null) {
       clearTimeout(streamEndTimerRef.current);
       streamEndTimerRef.current = null;
     }
     // eslint-disable-next-line react-hooks/exhaustive-deps
-  }, [chatId, client]);
+  }, [chatId, client, clearActivitySegment, clearPendingStreamWork]);
 
   useEffect(() => {
     if (hasPendingToolCalls) setIsStreaming(true);
@@ -296,54 +533,10 @@ export function useNanobotStream(
       if (ev.event === "delta") {
         if (suppressStreamUntilTurnEndRef.current) return;
         const chunk = typeof ev.text === "string" ? ev.text : "";
+        if (!chunk) return;
         setIsStreaming(true);
-        setMessages((prev) => {
-          const adopted = findActiveAssistantPlaceholder(prev);
-          const streamingAssistId = findStreamingAssistantId(prev);
-          let targetId: string;
-          let next: UIMessage[];
-
-          if (adopted) {
-            targetId = adopted;
-            next = prev;
-          } else if (streamingAssistId) {
-            targetId = streamingAssistId;
-            next = prev;
-          } else {
-            targetId = crypto.randomUUID();
-            next = [
-              ...prev,
-              {
-                id: targetId,
-                role: "assistant",
-                content: "",
-                isStreaming: true,
-                createdAt: Date.now(),
-              },
-            ];
-          }
-
-          buffer.current = { messageId: targetId };
-
-          const priorContent = next.find((m) => m.id === targetId)?.content ?? "";
-          const combined = priorContent + chunk;
-          return next.map((m) =>
-            m.id === targetId ? { ...m, content: combined, isStreaming: true } : m,
-          );
-        });
-        return;
-      }
-
-      if (ev.event === "stream_end") {
-        if (suppressStreamUntilTurnEndRef.current) {
-          buffer.current = null;
-          return;
-        }
-        // stream_end only means the text segment finished — the model may
-        // still be executing tools.  Do NOT reset isStreaming here; the
-        // definitive "turn is complete" signal is ``turn_end``.
-        if (!buffer.current) return;
-        buffer.current = null;
+        pendingStreamEventsRef.current.push({ kind: "delta", text: chunk });
+        schedulePendingStreamFlush();
         return;
       }
 
@@ -351,11 +544,23 @@ export function useNanobotStream(
         if (suppressStreamUntilTurnEndRef.current) return;
         const chunk = ev.text;
         if (!chunk) return;
-        setMessages((prev) => attachReasoningChunk(prev, chunk));
         setIsStreaming(true);
+        pendingStreamEventsRef.current.push({ kind: "reasoning", text: chunk });
+        schedulePendingStreamFlush();
         return;
       }
 
+      if (ev.event === "stream_end") {
+        flushPendingStreamEvents({ closeAnswerSegment: true });
+        if (suppressStreamUntilTurnEndRef.current) return;
+        // stream_end only means the text segment finished — the model may
+        // still be executing tools.  Do NOT reset isStreaming here; the
+        // definitive "turn is complete" signal is ``turn_end``.
+        return;
+      }
+
+      flushPendingStreamEvents();
+
       if (ev.event === "reasoning_end") {
         if (suppressStreamUntilTurnEndRef.current) return;
         setMessages((prev) => closeReasoningStream(prev));
@@ -393,6 +598,10 @@ export function useNanobotStream(
           if (typeof ev.latency_ms === "number" && ev.latency_ms >= 0) {
             finalized = stampLastAssistantLatency(finalized, Math.round(ev.latency_ms));
           }
+          buffer.current = null;
+          activeAssistantRef.current = null;
+          clearActivitySegment();
+          closedAssistantStreamIdsRef.current.clear();
           return finalized;
         });
         suppressStreamUntilTurnEndRef.current = false;
@@ -413,7 +622,9 @@ export function useNanobotStream(
         if (ev.kind === "reasoning") {
           const line = ev.text;
           if (!line) return;
-          setMessages((prev) => closeReasoningStream(attachReasoningChunk(prev, line)));
+          setMessages((prev) => closeReasoningStream(attachReasoningChunk(prev, line, {
+            ensure: ensureActivitySegmentId,
+          })));
           return;
         }
         // Intermediate agent breadcrumbs (tool-call hints, raw progress).
@@ -428,12 +639,24 @@ export function useNanobotStream(
               : [];
           if (lines.length === 0) return;
           setMessages((prev) => {
+            const segmentId = ensureActivitySegmentId();
             const last = prev[prev.length - 1];
-            if (last && last.kind === "trace" && !last.isStreaming) {
+            if (
+              last
+              && last.kind === "trace"
+              && !last.isStreaming
+              && (!last.activitySegmentId || last.activitySegmentId === segmentId)
+            ) {
+              const previousTraces = last.traces?.length
+                ? last.traces
+                : last.content
+                  ? [last.content]
+                  : [];
               const merged: UIMessage = {
                 ...last,
-                traces: [...(last.traces ?? [last.content]), ...lines],
+                traces: [...previousTraces, ...lines],
                 content: lines[lines.length - 1],
+                activitySegmentId: last.activitySegmentId ?? segmentId,
               };
               return [...prev.slice(0, -1), merged];
             }
@@ -445,6 +668,7 @@ export function useNanobotStream(
                 kind: "trace",
                 content: lines[lines.length - 1],
                 traces: lines,
+                activitySegmentId: segmentId,
                 createdAt: Date.now(),
               },
             ];
@@ -459,11 +683,12 @@ export function useNanobotStream(
 
         // A complete (non-streamed) assistant message. If a stream was in
         // flight, drop the placeholder so we don't render the text twice.
-        const activeId = buffer.current?.messageId;
-        buffer.current = null;
         // Do NOT reset isStreaming here — only ``turn_end`` signals that
         // the full turn (all tool calls + final text) is complete.
         setMessages((prev) => {
+          const activeId = buffer.current?.messageId;
+          buffer.current = null;
+          activeAssistantRef.current = null;
           const filtered = activeId ? prev.filter((m) => m.id !== activeId) : prev;
           const content = ev.text;
           const lat =
@@ -481,6 +706,46 @@ export function useNanobotStream(
         }
         return;
       }
+      if (ev.event === "file_edit") {
+        const edits = Array.isArray(ev.edits) ? ev.edits : [];
+        if (edits.length === 0) return;
+        setMessages((prev) => {
+          const last = prev[prev.length - 1];
+          let segmentId = fileEditSegmentRef.current;
+          if (!segmentId || !(last?.kind === "trace" && last.fileEdits?.length)) {
+            segmentId = detachedActivitySegmentId();
+            fileEditSegmentRef.current = segmentId;
+          }
+          if (
+            last
+            && last.kind === "trace"
+            && !last.isStreaming
+            && !!last.fileEdits?.length
+            && last.activitySegmentId === segmentId
+          ) {
+            const merged: UIMessage = {
+              ...last,
+              fileEdits: mergeFileEdits(last.fileEdits, edits),
+              activitySegmentId: last.activitySegmentId ?? segmentId,
+            };
+            return [...prev.slice(0, -1), merged];
+          }
+          return [
+            ...prev,
+            {
+              id: crypto.randomUUID(),
+              role: "tool",
+              kind: "trace",
+              content: "",
+              traces: [],
+              fileEdits: mergeFileEdits(undefined, edits),
+              activitySegmentId: segmentId,
+              createdAt: Date.now(),
+            },
+          ];
+        });
+        return;
+      }
       // ``attached`` / ``error`` frames aren't actionable here; the client
       // shell handles them separately.
     };
@@ -489,12 +754,26 @@ export function useNanobotStream(
     return () => {
       unsub();
       buffer.current = null;
+      activeAssistantRef.current = null;
+      closedAssistantStreamIdsRef.current.clear();
+      clearActivitySegment();
+      clearPendingStreamWork();
       if (streamEndTimerRef.current !== null) {
         clearTimeout(streamEndTimerRef.current);
         streamEndTimerRef.current = null;
       }
     };
-  }, [chatId, client, onTurnEnd]);
+  }, [
+    chatId,
+    client,
+    clearActivitySegment,
+    clearPendingStreamWork,
+    detachedActivitySegmentId,
+    ensureActivitySegmentId,
+    flushPendingStreamEvents,
+    onTurnEnd,
+    schedulePendingStreamFlush,
+  ]);
 
   const send = useCallback(
     (content: string, images?: SendImage[], options?: SendOptions) => {
@@ -504,17 +783,24 @@ export function useNanobotStream(
       // the image blocks via ``media`` paths.
       if (!hasImages && !content.trim()) return;
 
+      flushPendingStreamEvents();
       const previews = hasImages ? images!.map((i) => i.preview) : undefined;
-      setMessages((prev) => [
-        ...pruneReasoningOnlyPlaceholders(prev),
-        {
-          id: crypto.randomUUID(),
-          role: "user",
-          content,
-          createdAt: Date.now(),
-          ...(previews ? { images: previews } : {}),
-        },
-      ]);
+      setMessages((prev) => {
+        buffer.current = null;
+        activeAssistantRef.current = null;
+        closedAssistantStreamIdsRef.current.clear();
+        clearActivitySegment();
+        return [
+          ...pruneReasoningOnlyPlaceholders(prev),
+          {
+            id: crypto.randomUUID(),
+            role: "user",
+            content,
+            createdAt: Date.now(),
+            ...(previews ? { images: previews } : {}),
+          },
+        ];
+      });
       // Mark streaming immediately so the UI shows the loading indicator
       // right away, before the first delta arrives from the server.
       setIsStreaming(true);
@@ -525,18 +811,23 @@ export function useNanobotStream(
         client.sendMessage(chatId, content, wireMedia);
       }
     },
-    [chatId, client],
+    [chatId, clearActivitySegment, client, flushPendingStreamEvents],
   );
 
   const stop = useCallback(() => {
     if (!chatId) return;
+    flushPendingStreamEvents();
     setIsStreaming(false);
-    setMessages((prev) =>
-      prev.map((m) => (m.isStreaming ? { ...m, isStreaming: false } : m)),
-    );
+    setMessages((prev) => {
+      buffer.current = null;
+      activeAssistantRef.current = null;
+      closedAssistantStreamIdsRef.current.clear();
+      clearActivitySegment();
+      return prev.map((m) => (m.isStreaming ? { ...m, isStreaming: false } : m));
+    });
     suppressStreamUntilTurnEndRef.current = false;
     client.sendMessage(chatId, "/stop");
-  }, [chatId, client]);
+  }, [chatId, clearActivitySegment, client, flushPendingStreamEvents]);
 
   return {
     messages,

webui/src/hooks/useTheme.ts

@@ -1,7 +1,16 @@
-import { useCallback, useEffect, useState } from "react";
+import {
+  createContext,
+  createElement,
+  useCallback,
+  useContext,
+  useEffect,
+  useState,
+  type ReactNode,
+} from "react";
 
 type Theme = "light" | "dark";
 const STORAGE_KEY = "nanobot-webui.theme";
+const ThemeContext = createContext<Theme>("light");
 
 function readStored(): Theme | null {
   try {
@@ -18,7 +27,11 @@ function applyTheme(theme: Theme): void {
   else root.classList.remove("dark");
 }
 
-export function useTheme(): { theme: Theme; toggle: () => void; setTheme: (t: Theme) => void } {
+export function useTheme(): {
+  theme: Theme;
+  toggle: () => void;
+  setTheme: (t: Theme) => void;
+} {
   const [theme, setThemeState] = useState<Theme>(() => {
     const stored = readStored();
     if (stored) return stored;
@@ -46,3 +59,11 @@ export function useTheme(): { theme: Theme; toggle: () => void; setTheme: (t: Th
   );
   return { theme, toggle, setTheme };
 }
+
+export function ThemeProvider({ theme, children }: { theme: Theme; children: ReactNode }) {
+  return createElement(ThemeContext.Provider, { value: theme }, children);
+}
+
+export function useThemeValue(): Theme {
+  return useContext(ThemeContext);
+}

webui/src/i18n/locales/en/common.json

@@ -335,7 +335,8 @@
         "io": "Couldn't read this file"
       }
     },
-    "scrollToBottom": "Scroll to bottom"
+    "scrollToBottom": "Scroll to bottom",
+    "loadEarlier": "Load earlier messages"
   },
   "message": {
     "streaming": "streaming",

webui/src/i18n/locales/es/common.json

@@ -303,7 +303,8 @@
       },
       "goalStateCloseAria": "Cerrar objetivo"
     },
-    "scrollToBottom": "Desplazarse al final"
+    "scrollToBottom": "Desplazarse al final",
+    "loadEarlier": "Cargar mensajes anteriores"
   },
   "message": {
     "streaming": "transmitiendo",

webui/src/i18n/locales/fr/common.json

@@ -303,7 +303,8 @@
       },
       "goalStateCloseAria": "Fermer l’objectif"
     },
-    "scrollToBottom": "Faire défiler vers le bas"
+    "scrollToBottom": "Faire défiler vers le bas",
+    "loadEarlier": "Charger les messages précédents"
   },
   "message": {
     "streaming": "en cours de génération",

webui/src/i18n/locales/id/common.json

@@ -303,7 +303,8 @@
       },
       "goalStateCloseAria": "Tutup tujuan"
     },
-    "scrollToBottom": "Gulir ke bawah"
+    "scrollToBottom": "Gulir ke bawah",
+    "loadEarlier": "Muat pesan sebelumnya"
   },
   "message": {
     "streaming": "sedang mengalir",

webui/src/i18n/locales/ja/common.json

@@ -303,7 +303,8 @@
       },
       "goalStateCloseAria": "目標を閉じる"
     },
-    "scrollToBottom": "一番下へスクロール"
+    "scrollToBottom": "一番下へスクロール",
+    "loadEarlier": "以前のメッセージを読み込む"
   },
   "message": {
     "streaming": "生成中",

webui/src/i18n/locales/ko/common.json

@@ -303,7 +303,8 @@
       },
       "goalStateCloseAria": "목표 닫기"
     },
-    "scrollToBottom": "맨 아래로 스크롤"
+    "scrollToBottom": "맨 아래로 스크롤",
+    "loadEarlier": "이전 메시지 불러오기"
   },
   "message": {
     "streaming": "생성 중",

webui/src/i18n/locales/vi/common.json

@@ -303,7 +303,8 @@
       },
       "goalStateCloseAria": "Đóng mục tiêu"
     },
-    "scrollToBottom": "Cuộn xuống cuối"
+    "scrollToBottom": "Cuộn xuống cuối",
+    "loadEarlier": "Tải tin nhắn trước đó"
   },
   "message": {
     "streaming": "đang truyền",

webui/src/i18n/locales/zh-CN/common.json

@@ -323,7 +323,8 @@
       },
       "goalStateCloseAria": "关闭目标"
     },
-    "scrollToBottom": "滚动到底部"
+    "scrollToBottom": "滚动到底部",
+    "loadEarlier": "加载更早消息"
   },
   "message": {
     "streaming": "流式输出中",

webui/src/i18n/locales/zh-TW/common.json

@@ -303,7 +303,8 @@
       },
       "goalStateCloseAria": "關閉目標"
     },
-    "scrollToBottom": "捲動到底部"
+    "scrollToBottom": "捲動到底部",
+    "loadEarlier": "載入更早訊息"
   },
   "message": {
     "streaming": "串流輸出中",

webui/src/lib/format.ts

@@ -1,10 +1,35 @@
 import i18n, { currentLocale } from "@/i18n";
 
+const LOW_INFORMATION_TITLE_PREVIEWS = new Set([
+  "hi",
+  "hello",
+  "hey",
+  "hello nano",
+  "hello nanobot",
+  "hi nano",
+  "hi nanobot",
+  "你好",
+  "您好",
+  "嗨",
+  "哈喽",
+  "哈啰",
+  "在吗",
+]);
+
+function isLowInformationTitlePreview(text: string): boolean {
+  const normalized = text.toLowerCase().replace(/[.!?。！？~～\s]+$/g, "").trim();
+  return (
+    normalized.startsWith("/") ||
+    LOW_INFORMATION_TITLE_PREVIEWS.has(normalized)
+  );
+}
+
 /** Truncate the first user message into a chat title. */
 export function deriveTitle(preview: string | undefined, fallback: string): string {
   if (!preview) return fallback;
   const oneLine = preview.replace(/\s+/g, " ").trim();
   if (!oneLine) return fallback;
+  if (isLowInformationTitlePreview(oneLine)) return fallback;
   return oneLine.length > 60 ? `${oneLine.slice(0, 57)}…` : oneLine;
 }
 

webui/src/lib/nanobot-client.ts

@@ -54,7 +54,8 @@ type Unsubscribe = () => void;
 type EventHandler = (ev: InboundEvent) => void;
 type StatusHandler = (status: ConnectionStatus) => void;
 type RuntimeModelHandler = (modelName: string | null, modelPreset?: string | null) => void;
-type SessionUpdateHandler = (chatId: string) => void;
+type SessionUpdateScope = "metadata" | "thread" | string;
+type SessionUpdateHandler = (chatId: string, scope?: SessionUpdateScope) => void;
 
 /** Structured connection-level errors surfaced to the UI.
  *
@@ -364,7 +365,7 @@ export class NanobotClient {
     }
 
     if (parsed.event === "session_updated") {
-      this.emitSessionUpdate(parsed.chat_id);
+      this.emitSessionUpdate(parsed.chat_id, parsed.scope);
       return;
     }
 
@@ -382,9 +383,9 @@ export class NanobotClient {
     }
   }
 
-  private emitSessionUpdate(chatId: string): void {
+  private emitSessionUpdate(chatId: string, scope?: SessionUpdateScope): void {
     for (const handler of this.sessionUpdateHandlers) {
-      handler(chatId);
+      handler(chatId, scope);
     }
   }
 

webui/src/lib/types.ts

@@ -40,6 +40,10 @@ export interface UIMessage {
   /** For trace rows: each individual hint line, so consecutive hints can
    * render as a single collapsible group. */
   traces?: string[];
+  /** Activity rows: explicit file edits emitted by edit tools. */
+  fileEdits?: UIFileEdit[];
+  /** Activity rows created during the same agent phase share one collapsible block. */
+  activitySegmentId?: string;
   /** User turn: optimistic blob URLs for preview. Replay: placeholder chips. */
   images?: UIImage[];
   /** Signed or local UI-renderable media attachments. */
@@ -80,6 +84,20 @@ export interface ToolProgressEvent {
   embeds?: unknown[];
 }
 
+export interface UIFileEdit {
+  version?: number;
+  call_id: string;
+  tool: string;
+  path: string;
+  phase?: "start" | "end" | "error" | string;
+  added: number;
+  deleted: number;
+  approximate?: boolean;
+  status: "editing" | "done" | "error";
+  binary?: boolean;
+  error?: string;
+}
+
 export interface ChatSummary {
   /** Server-side session key, e.g. ``websocket:abcd-...``. */
   key: string;
@@ -110,6 +128,7 @@ export interface SettingsPayload {
     name: string;
     label: string;
     configured: boolean;
+    api_key_required?: boolean;
     api_key_hint?: string | null;
     api_base?: string | null;
     default_api_base?: string | null;
@@ -182,6 +201,11 @@ export type InboundEvent =
       /** Optional structured payload on progress frames (channel-specific). */
       agent_ui?: AgentUIBlob;
     }
+  | {
+      event: "file_edit";
+      chat_id: string;
+      edits: UIFileEdit[];
+    }
   | {
       event: "delta";
       chat_id: string;
@@ -229,7 +253,7 @@ export type InboundEvent =
       chat_id: string;
       goal_state: GoalStateWsPayload;
     }
-  | { event: "session_updated"; chat_id: string }
+  | { event: "session_updated"; chat_id: string; scope?: "metadata" | "thread" | string }
   | { event: "error"; chat_id?: string; detail?: string };
 
 /** Base64-encoded image attached to an outbound ``message`` envelope.

webui/src/main.tsx

@@ -1,4 +1,3 @@
-import React from "react";
 import ReactDOM from "react-dom/client";
 
 import App from "./App";

webui/src/tests/agent-activity-cluster.test.tsx

@@ -0,0 +1,336 @@
+import { act, fireEvent, render, screen, waitFor } from "@testing-library/react";
+import { describe, expect, it } from "vitest";
+
+import { AgentActivityCluster } from "@/components/thread/AgentActivityCluster";
+import type { UIMessage } from "@/lib/types";
+
+function activityMessages(extraReasoning = "", extraTool?: UIMessage): UIMessage[] {
+  const rows: UIMessage[] = [
+    {
+      id: "r1",
+      role: "assistant",
+      content: "",
+      reasoning: `thinking${extraReasoning}`,
+      reasoningStreaming: true,
+      isStreaming: true,
+      createdAt: 1,
+    },
+    {
+      id: "t1",
+      role: "tool",
+      kind: "trace",
+      content: "search()",
+      traces: ["search()"],
+      createdAt: 2,
+    },
+  ];
+  if (extraTool) rows.push(extraTool);
+  return rows;
+}
+
+function installAnimationFrameQueue() {
+  const originalRequest = window.requestAnimationFrame;
+  const originalCancel = window.cancelAnimationFrame;
+  const callbacks = new Map<number, FrameRequestCallback>();
+  let nextId = 1;
+
+  window.requestAnimationFrame = ((callback: FrameRequestCallback) => {
+    const id = nextId;
+    nextId += 1;
+    callbacks.set(id, callback);
+    return id;
+  }) as typeof window.requestAnimationFrame;
+  window.cancelAnimationFrame = ((id: number) => {
+    callbacks.delete(id);
+  }) as typeof window.cancelAnimationFrame;
+
+  return {
+    flush() {
+      const pending = Array.from(callbacks.entries());
+      callbacks.clear();
+      for (const [, callback] of pending) callback(0);
+    },
+    restore() {
+      window.requestAnimationFrame = originalRequest;
+      window.cancelAnimationFrame = originalCancel;
+    },
+  };
+}
+
+function setScrollGeometry(
+  element: HTMLElement,
+  geometry: { scrollHeight: number; clientHeight: number; scrollTop?: number },
+) {
+  Object.defineProperties(element, {
+    scrollHeight: { configurable: true, value: geometry.scrollHeight },
+    clientHeight: { configurable: true, value: geometry.clientHeight },
+    scrollTop: {
+      configurable: true,
+      value: geometry.scrollTop ?? element.scrollTop,
+      writable: true,
+    },
+  });
+}
+
+function installReducedMotion() {
+  const original = window.matchMedia;
+  Object.defineProperty(window, "matchMedia", {
+    configurable: true,
+    value: () => ({
+      matches: true,
+      media: "(prefers-reduced-motion: reduce)",
+      addEventListener: () => {},
+      removeEventListener: () => {},
+    }),
+  });
+  return () => {
+    Object.defineProperty(window, "matchMedia", {
+      configurable: true,
+      value: original,
+    });
+  };
+}
+
+describe("AgentActivityCluster", () => {
+  it("jumps to the latest activity when opened", () => {
+    const raf = installAnimationFrameQueue();
+    try {
+      render(
+        <AgentActivityCluster
+          messages={activityMessages()}
+          isTurnStreaming
+          hasBodyBelow={false}
+        />,
+      );
+
+      fireEvent.click(screen.getByRole("button", { name: /working/i }));
+      const scrollport = screen.getByTestId("agent-activity-scroll");
+      setScrollGeometry(scrollport, {
+        scrollHeight: 1000,
+        clientHeight: 120,
+        scrollTop: 0,
+      });
+
+      act(() => {
+        raf.flush();
+      });
+
+      expect(scrollport.scrollTop).toBe(880);
+    } finally {
+      raf.restore();
+    }
+  });
+
+  it("follows new reasoning and tool activity while the user is at the bottom", () => {
+    const raf = installAnimationFrameQueue();
+    try {
+      const { rerender } = render(
+        <AgentActivityCluster
+          messages={activityMessages()}
+          isTurnStreaming
+          hasBodyBelow={false}
+        />,
+      );
+
+      fireEvent.click(screen.getByRole("button", { name: /working/i }));
+      const scrollport = screen.getByTestId("agent-activity-scroll");
+      setScrollGeometry(scrollport, {
+        scrollHeight: 1000,
+        clientHeight: 120,
+        scrollTop: 0,
+      });
+      act(() => {
+        raf.flush();
+      });
+
+      rerender(
+        <AgentActivityCluster
+          messages={activityMessages(" with more detail", {
+            id: "t2",
+            role: "tool",
+            kind: "trace",
+            content: "open_browser()",
+            traces: ["open_browser()"],
+            createdAt: 3,
+          })}
+          isTurnStreaming
+          hasBodyBelow={false}
+        />,
+      );
+      setScrollGeometry(scrollport, {
+        scrollHeight: 1500,
+        clientHeight: 120,
+        scrollTop: scrollport.scrollTop,
+      });
+
+      act(() => {
+        raf.flush();
+      });
+
+      expect(scrollport.scrollTop).toBe(1380);
+    } finally {
+      raf.restore();
+    }
+  });
+
+  it("does not pull the user down after they scroll up inside the activity pane", () => {
+    const raf = installAnimationFrameQueue();
+    try {
+      const { rerender } = render(
+        <AgentActivityCluster
+          messages={activityMessages()}
+          isTurnStreaming
+          hasBodyBelow={false}
+        />,
+      );
+
+      fireEvent.click(screen.getByRole("button", { name: /working/i }));
+      const scrollport = screen.getByTestId("agent-activity-scroll");
+      setScrollGeometry(scrollport, {
+        scrollHeight: 1000,
+        clientHeight: 120,
+        scrollTop: 0,
+      });
+      act(() => {
+        raf.flush();
+      });
+
+      scrollport.scrollTop = 100;
+      fireEvent.scroll(scrollport);
+
+      rerender(
+        <AgentActivityCluster
+          messages={activityMessages(" still streaming")}
+          isTurnStreaming
+          hasBodyBelow={false}
+        />,
+      );
+      setScrollGeometry(scrollport, {
+        scrollHeight: 1500,
+        clientHeight: 120,
+        scrollTop: scrollport.scrollTop,
+      });
+
+      act(() => {
+        raf.flush();
+      });
+
+      expect(scrollport.scrollTop).toBe(100);
+    } finally {
+      raf.restore();
+    }
+  });
+
+  it("renders file edit totals and a compact expanded file list", async () => {
+    const restoreMotion = installReducedMotion();
+    try {
+      render(
+        <AgentActivityCluster
+          messages={activityMessages("", {
+            id: "t2",
+            role: "tool",
+            kind: "trace",
+            content: "edit_file()",
+            traces: ["edit_file()"],
+            fileEdits: [{
+              call_id: "call-edit",
+              tool: "edit_file",
+              path: "src/app.tsx",
+              phase: "end",
+              added: 12,
+              deleted: 3,
+              approximate: false,
+              status: "done",
+            }],
+            createdAt: 3,
+          })}
+          isTurnStreaming={false}
+          hasBodyBelow={false}
+        />,
+      );
+
+      expect(screen.getByRole("button", { name: /edited app\.tsx/i })).toBeInTheDocument();
+      fireEvent.click(screen.getByRole("button", { name: /edited app\.tsx/i }));
+
+      expect(screen.queryByText("Edited files")).not.toBeInTheDocument();
+      expect(screen.queryByText("Edited")).not.toBeInTheDocument();
+      const fileRef = screen.getByTestId("activity-file-reference");
+      expect(fileRef).toHaveTextContent("src/app.tsx");
+      expect(fileRef).toHaveAttribute("aria-label", "src/app.tsx");
+      await waitFor(() => {
+        expect(screen.getAllByText("+12").length).toBeGreaterThan(0);
+        expect(screen.getAllByText("-3").length).toBeGreaterThan(0);
+      });
+    } finally {
+      restoreMotion();
+    }
+  });
+
+  it("merges repeated edits for the same path and lets successful edits win over failures", async () => {
+    const restoreMotion = installReducedMotion();
+    try {
+      render(
+        <AgentActivityCluster
+          messages={activityMessages("", {
+            id: "t2",
+            role: "tool",
+            kind: "trace",
+            content: "edit_file()",
+            traces: ["edit_file()"],
+            fileEdits: [
+              {
+                call_id: "call-edit-1",
+                tool: "edit_file",
+                path: "minecraft-fps/index.html",
+                phase: "end",
+                added: 2,
+                deleted: 1,
+                approximate: false,
+                status: "done",
+              },
+              {
+                call_id: "call-edit-2",
+                tool: "edit_file",
+                path: "minecraft-fps/index.html",
+                phase: "error",
+                added: 0,
+                deleted: 0,
+                approximate: false,
+                status: "error",
+                error: "patch failed",
+              },
+              {
+                call_id: "call-edit-3",
+                tool: "edit_file",
+                path: "minecraft-fps/index.html",
+                phase: "end",
+                added: 6,
+                deleted: 6,
+                approximate: false,
+                status: "done",
+              },
+            ],
+            createdAt: 3,
+          })}
+          isTurnStreaming={false}
+          hasBodyBelow={false}
+        />,
+      );
+
+      expect(screen.getByRole("button", { name: /edited index\.html/i })).toBeInTheDocument();
+      expect(screen.queryByRole("button", { name: /failed index\.html/i })).not.toBeInTheDocument();
+      fireEvent.click(screen.getByRole("button", { name: /edited index\.html/i }));
+
+      const fileRefs = screen.getAllByTestId("activity-file-reference");
+      expect(fileRefs).toHaveLength(1);
+      expect(fileRefs[0]).toHaveTextContent("minecraft-fps/index.html");
+      expect(screen.queryByText("Failed")).not.toBeInTheDocument();
+      await waitFor(() => {
+        expect(screen.getAllByText("+8").length).toBeGreaterThan(0);
+        expect(screen.getAllByText("-7").length).toBeGreaterThan(0);
+      });
+    } finally {
+      restoreMotion();
+    }
+  });
+});

webui/src/tests/app-layout.test.tsx

@@ -1,5 +1,5 @@
-import { fireEvent, render, screen, waitFor, within } from "@testing-library/react";
-import { beforeEach, describe, expect, it, vi } from "vitest";
+import { act, fireEvent, render, screen, waitFor, within } from "@testing-library/react";
+import { afterEach, beforeEach, describe, expect, it, vi } from "vitest";
 
 import type { ChatSummary } from "@/lib/types";
 
@@ -8,6 +8,7 @@ const refreshSpy = vi.fn();
 const createChatSpy = vi.fn().mockResolvedValue("chat-1");
 const deleteChatSpy = vi.fn();
 const toggleThemeSpy = vi.fn();
+const updateUrlSpy = vi.fn();
 let mockSessions: ChatSummary[] = [];
 
 vi.mock("@/hooks/useSessions", async (importOriginal) => {
@@ -32,12 +33,18 @@ vi.mock("@/hooks/useSessions", async (importOriginal) => {
   };
 });
 
-vi.mock("@/hooks/useTheme", () => ({
-  useTheme: () => ({
-    theme: "light" as const,
-    toggle: toggleThemeSpy,
-  }),
-}));
+vi.mock("@/hooks/useTheme", async () => {
+  const React = await import("react");
+  return {
+    ThemeProvider: ({ children }: { children: React.ReactNode }) =>
+      React.createElement(React.Fragment, null, children),
+    useTheme: () => ({
+      theme: "light" as const,
+      toggle: toggleThemeSpy,
+    }),
+    useThemeValue: () => "light" as const,
+  };
+});
 
 vi.mock("@/lib/bootstrap", () => ({
   fetchBootstrap: vi.fn().mockResolvedValue({
@@ -64,22 +71,30 @@ vi.mock("@/lib/nanobot-client", () => {
     newChat = vi.fn();
     attach = vi.fn();
     close = vi.fn();
-    updateUrl = vi.fn();
+    updateUrl = updateUrlSpy;
   }
 
   return { NanobotClient: MockClient };
 });
 
+import { deriveWsUrl, fetchBootstrap } from "@/lib/bootstrap";
 import App from "@/App";
 
 describe("App layout", () => {
   beforeEach(() => {
     mockSessions = [];
     connectSpy.mockClear();
+    updateUrlSpy.mockClear();
     refreshSpy.mockReset();
     createChatSpy.mockClear();
     deleteChatSpy.mockReset();
     toggleThemeSpy.mockReset();
+    vi.mocked(fetchBootstrap).mockReset().mockResolvedValue({
+      token: "tok",
+      ws_path: "/",
+      expires_in: 300,
+    });
+    vi.mocked(deriveWsUrl).mockReset().mockReturnValue("ws://test");
     vi.stubGlobal(
       "fetch",
       vi.fn().mockResolvedValue({
@@ -89,6 +104,10 @@ describe("App layout", () => {
     );
   });
 
+  afterEach(() => {
+    vi.useRealTimers();
+  });
+
   it("keeps sidebar layout out of the main thread width contract", async () => {
     const { container } = render(<App />);
 
@@ -192,8 +211,52 @@ describe("App layout", () => {
                   name: "openrouter",
                   label: "OpenRouter",
                   configured: false,
+                  api_key_required: true,
                   default_api_base: "https://openrouter.ai/api/v1",
                 },
+                {
+                  name: "azure_openai",
+                  label: "Azure OpenAI",
+                  configured: false,
+                  api_key_required: true,
+                },
+                {
+                  name: "huggingface",
+                  label: "Hugging Face",
+                  configured: false,
+                  api_key_required: true,
+                },
+                {
+                  name: "siliconflow",
+                  label: "SiliconFlow",
+                  configured: false,
+                  api_key_required: true,
+                },
+                {
+                  name: "volcengine",
+                  label: "VolcEngine",
+                  configured: false,
+                  api_key_required: true,
+                },
+                {
+                  name: "byteplus",
+                  label: "BytePlus",
+                  configured: false,
+                  api_key_required: true,
+                },
+                {
+                  name: "qianfan",
+                  label: "Qianfan",
+                  configured: false,
+                  api_key_required: true,
+                },
+                {
+                  name: "atomic_chat",
+                  label: "Atomic Chat",
+                  configured: false,
+                  api_key_required: false,
+                  default_api_base: "http://localhost:1337/v1",
+                },
               ],
               web_search: {
                 provider: "brave",
@@ -248,6 +311,9 @@ describe("App layout", () => {
     fireEvent.click(screen.getByText("OpenAI"));
     expect(screen.getByText("open••••-key")).toBeInTheDocument();
     expect(screen.queryByDisplayValue("unsaved-openai-key")).not.toBeInTheDocument();
+    fireEvent.click(screen.getByText("Atomic Chat"));
+    expect(screen.getByDisplayValue("http://localhost:1337/v1")).toBeInTheDocument();
+    expect(screen.getByRole("button", { name: "Save" })).toBeEnabled();
 
     fireEvent.click(screen.getByRole("tab", { name: "Web Search" }));
     expect(screen.getByText("Search provider")).toBeInTheDocument();
@@ -426,4 +492,36 @@ describe("App layout", () => {
 
     expect(within(sidebar).getByText("Existing chat")).toBeInTheDocument();
   });
+
+  it("refreshes the bootstrap token before REST settings auth expires", async () => {
+    vi.useFakeTimers();
+    vi.mocked(fetchBootstrap)
+      .mockResolvedValueOnce({
+        token: "tok-1",
+        ws_path: "/",
+        expires_in: 30,
+      })
+      .mockResolvedValueOnce({
+        token: "tok-2",
+        ws_path: "/",
+        expires_in: 300,
+      });
+    vi.mocked(deriveWsUrl).mockImplementation(
+      (_wsPath: string, token: string) => `ws://test?token=${token}`,
+    );
+
+    const { unmount } = render(<App />);
+    await act(async () => {});
+
+    expect(connectSpy).toHaveBeenCalled();
+    expect(fetchBootstrap).toHaveBeenCalledTimes(1);
+
+    await act(async () => {
+      await vi.advanceTimersByTimeAsync(15_000);
+    });
+
+    expect(fetchBootstrap).toHaveBeenCalledTimes(2);
+    expect(updateUrlSpy).toHaveBeenCalledWith("ws://test?token=tok-2");
+    unmount();
+  });
 });

webui/src/tests/code-block.test.tsx

@@ -0,0 +1,104 @@
+import { act, render, screen } from "@testing-library/react";
+import { describe, expect, it, vi } from "vitest";
+
+import { CodeBlock } from "@/components/CodeBlock";
+import { ThemeProvider } from "@/hooks/useTheme";
+
+const mockedStyles = vi.hoisted(() => ({
+  dark: { pre: { background: "#111" } },
+  light: { pre: { background: "#fff" } },
+}));
+
+vi.mock("react-syntax-highlighter/dist/esm/prism-async-light", () => ({
+  default: ({
+    children,
+    style,
+  }: {
+    children: string;
+    style: Record<string, unknown>;
+  }) => (
+    <pre
+      data-testid="highlighted-code"
+      data-theme={style === mockedStyles.dark ? "dark" : "light"}
+    >
+      <code>{children}</code>
+    </pre>
+  ),
+}));
+
+vi.mock("react-syntax-highlighter/dist/esm/styles/prism/one-dark", () => ({
+  default: mockedStyles.dark,
+}));
+
+vi.mock("react-syntax-highlighter/dist/esm/styles/prism/one-light", () => ({
+  default: mockedStyles.light,
+}));
+
+describe("CodeBlock", () => {
+  it("renders plain code without mounting the highlighter when highlighting is disabled", () => {
+    render(
+      <ThemeProvider theme="dark">
+        <CodeBlock language="ts" code="const value = 1;" highlight={false} />
+      </ThemeProvider>,
+    );
+
+    expect(screen.queryByTestId("highlighted-code")).not.toBeInTheDocument();
+    expect(screen.getByText("const value = 1;")).toBeInTheDocument();
+    expect(screen.getByText("ts")).toBeInTheDocument();
+  });
+
+  it("reads theme from context without creating per-block observers", async () => {
+    const originalMutationObserver = globalThis.MutationObserver;
+    const observer = vi.fn();
+    class MockMutationObserver {
+      constructor(callback: MutationCallback) {
+        observer(callback);
+      }
+
+      observe = vi.fn();
+
+      disconnect = vi.fn();
+
+      takeRecords() {
+        return [];
+      }
+    }
+    vi.stubGlobal("MutationObserver", MockMutationObserver);
+
+    try {
+      const { rerender } = render(
+        <ThemeProvider theme="dark">
+          <CodeBlock language="ts" code="const value = 1;" />
+        </ThemeProvider>,
+      );
+
+      await act(async () => {
+        await Promise.resolve();
+        await Promise.resolve();
+      });
+
+      expect(screen.getByTestId("highlighted-code")).toHaveAttribute(
+        "data-theme",
+        "dark",
+      );
+
+      rerender(
+        <ThemeProvider theme="light">
+          <CodeBlock language="ts" code="const value = 1;" />
+        </ThemeProvider>,
+      );
+
+      await act(async () => {
+        await Promise.resolve();
+      });
+
+      expect(screen.getByTestId("highlighted-code")).toHaveAttribute(
+        "data-theme",
+        "light",
+      );
+      expect(observer).not.toHaveBeenCalled();
+    } finally {
+      vi.stubGlobal("MutationObserver", originalMutationObserver);
+    }
+  });
+});

webui/src/tests/markdown-text.test.tsx

@@ -0,0 +1,82 @@
+import { act, render, screen } from "@testing-library/react";
+import { describe, expect, it, vi } from "vitest";
+
+import { MarkdownText } from "@/components/MarkdownText";
+
+const rendererSpy = vi.hoisted(() => vi.fn());
+
+vi.mock("@/components/MarkdownTextRenderer", () => ({
+  default: ({
+    children,
+    highlightCode,
+  }: {
+    children: string;
+    highlightCode?: boolean;
+  }) => {
+    rendererSpy({ children, highlightCode });
+    return (
+      <div
+        data-testid="markdown-renderer"
+        data-highlight-code={String(highlightCode)}
+      >
+        {children}
+      </div>
+    );
+  },
+}));
+
+describe("MarkdownText", () => {
+  it("throttles streaming markdown commits and flushes before final highlighting", async () => {
+    rendererSpy.mockClear();
+    vi.useFakeTimers();
+    try {
+      const { rerender } = render(
+        <MarkdownText streaming>hello</MarkdownText>,
+      );
+
+      await act(async () => {
+        await Promise.resolve();
+        await Promise.resolve();
+      });
+
+      expect(screen.getByTestId("markdown-renderer")).toHaveTextContent("hello");
+      expect(screen.getByTestId("markdown-renderer")).toHaveAttribute(
+        "data-highlight-code",
+        "false",
+      );
+      expect(rendererSpy).toHaveBeenCalledTimes(1);
+
+      rerender(<MarkdownText streaming>hello world</MarkdownText>);
+      expect(screen.getByTestId("markdown-renderer")).toHaveTextContent("hello");
+      expect(rendererSpy).toHaveBeenCalledTimes(1);
+
+      act(() => {
+        vi.advanceTimersByTime(79);
+      });
+      expect(screen.getByTestId("markdown-renderer")).toHaveTextContent("hello");
+      expect(rendererSpy).toHaveBeenCalledTimes(1);
+
+      act(() => {
+        vi.advanceTimersByTime(1);
+      });
+      await act(async () => {
+        await Promise.resolve();
+      });
+
+      expect(screen.getByTestId("markdown-renderer")).toHaveTextContent("hello world");
+      expect(rendererSpy).toHaveBeenCalledTimes(2);
+
+      rerender(<MarkdownText streaming>hello world!!!</MarkdownText>);
+      expect(screen.getByTestId("markdown-renderer")).toHaveTextContent("hello world");
+
+      rerender(<MarkdownText>hello world!!!</MarkdownText>);
+      expect(screen.getByTestId("markdown-renderer")).toHaveTextContent("hello world!!!");
+      expect(screen.getByTestId("markdown-renderer")).toHaveAttribute(
+        "data-highlight-code",
+        "true",
+      );
+    } finally {
+      vi.useRealTimers();
+    }
+  });
+});

webui/src/tests/message-bubble.test.tsx

@@ -1,4 +1,4 @@
-import { fireEvent, render, screen, waitFor } from "@testing-library/react";
+import { act, fireEvent, render, screen, waitFor } from "@testing-library/react";
 import { describe, expect, it, vi } from "vitest";
 
 import { MessageBubble } from "@/components/MessageBubble";
@@ -131,7 +131,9 @@ describe("MessageBubble", () => {
 
     expect(screen.getByText("Thinking…")).toBeInTheDocument();
     expect(screen.getByText(/Step 1: parse intent\./)).toBeInTheDocument();
-    expect(container.querySelector(".reasoning-sheen-stripe")).toBeInTheDocument();
+    expect(container.querySelector(".reasoning-sheen-stripe")).not.toBeInTheDocument();
+    expect(screen.getByText("Thinking…")).toHaveClass("streaming-text-sheen");
+    expect(screen.getByText("Thinking…")).toHaveAttribute("data-sheen-text", "Thinking…");
     expect(screen.getByRole("button", { name: /thinking/i }).parentElement).not.toHaveClass("mb-2");
   });
 
@@ -177,6 +179,47 @@ describe("MessageBubble", () => {
     expect(screen.getByText("Body line.")).toBeInTheDocument();
   });
 
+  it("renders inline file paths as compact file references", async () => {
+    await import("@/components/MarkdownTextRenderer");
+    const message: UIMessage = {
+      id: "a-file-path",
+      role: "assistant",
+      content:
+        "改动在 `webui/src/components/MarkdownTextRenderer.tsx` 和 `/Users/renxubin/.nanobot/workspace/minecraft-fps/index.html`。",
+      createdAt: Date.now(),
+    };
+
+    try {
+      render(<MessageBubble message={message} />);
+
+      const references = await screen.findAllByTestId("inline-file-path");
+      expect(references).toHaveLength(2);
+      expect(references[0].parentElement).not.toHaveClass("translate-y-[0.08em]");
+      expect(references[0].parentElement).toHaveClass("align-[0.14em]");
+      expect(references[0]).toHaveTextContent("MarkdownTextRenderer.tsx");
+      expect(references[0]).not.toHaveTextContent("webui/src/components");
+      expect(screen.getByText("index.html")).toBeInTheDocument();
+      expect(references[1]).not.toHaveTextContent("/Users/renxubin");
+      expect(references[1]).not.toHaveAttribute("title");
+      expect(references[1]).toHaveAttribute(
+        "aria-label",
+        "/Users/renxubin/.nanobot/workspace/minecraft-fps/index.html",
+      );
+
+      vi.useFakeTimers();
+      fireEvent.pointerMove(references[1].parentElement!);
+      await act(async () => {
+        vi.advanceTimersByTime(500);
+      });
+      const tooltip = screen.getByRole("tooltip");
+      expect(tooltip).toHaveTextContent(
+        "/Users/renxubin/.nanobot/workspace/minecraft-fps/index.html",
+      );
+    } finally {
+      vi.useRealTimers();
+    }
+  });
+
   it("renders assistant image media as a larger generated result", () => {
     const message: UIMessage = {
       id: "a-image",

webui/src/tests/nanobot-client.test.ts

@@ -233,9 +233,13 @@ describe("NanobotClient", () => {
     client.connect();
     lastSocket().fakeOpen();
 
-    lastSocket().fakeMessage({ event: "session_updated", chat_id: "chat-title" });
+    lastSocket().fakeMessage({
+      event: "session_updated",
+      chat_id: "chat-title",
+      scope: "metadata",
+    });
 
-    expect(globalHandler).toHaveBeenCalledWith("chat-title");
+    expect(globalHandler).toHaveBeenCalledWith("chat-title", "metadata");
     expect(chatHandler).not.toHaveBeenCalled();
   });
 

webui/src/tests/thread-messages.test.tsx

@@ -1,7 +1,11 @@
 import { render, screen } from "@testing-library/react";
 import { describe, expect, it } from "vitest";
 
-import { ThreadMessages } from "@/components/thread/ThreadMessages";
+import {
+  assistantCopyFlags,
+  buildDisplayUnits,
+  ThreadMessages,
+} from "@/components/thread/ThreadMessages";
 import type { UIMessage } from "@/lib/types";
 
 describe("ThreadMessages", () => {
@@ -51,6 +55,206 @@ describe("ThreadMessages", () => {
     expect(rows[1]).toHaveClass("mt-4");
   });
 
+  it("starts a new activity cluster when the activity segment changes", () => {
+    const messages: UIMessage[] = [
+      {
+        id: "r1",
+        role: "assistant",
+        content: "",
+        reasoning: "first pass",
+        activitySegmentId: "seg-1",
+        createdAt: 1,
+      },
+      {
+        id: "t1",
+        role: "tool",
+        kind: "trace",
+        content: "edit_file()",
+        traces: ["edit_file()"],
+        fileEdits: [{
+          call_id: "call-edit",
+          tool: "edit_file",
+          path: "foo.txt",
+          phase: "end",
+          added: 2,
+          deleted: 1,
+          status: "done",
+        }],
+        activitySegmentId: "seg-1",
+        createdAt: 2,
+      },
+      {
+        id: "r2",
+        role: "assistant",
+        content: "",
+        reasoning: "second pass",
+        activitySegmentId: "seg-2",
+        createdAt: 3,
+      },
+    ];
+
+    const units = buildDisplayUnits(messages);
+
+    expect(units).toHaveLength(2);
+    expect(units[0].type === "cluster" ? units[0].messages.map((m) => m.id) : []).toEqual([
+      "r1",
+      "t1",
+    ]);
+    expect(units[1].type === "cluster" ? units[1].messages.map((m) => m.id) : []).toEqual([
+      "r2",
+    ]);
+  });
+
+  it("does not split ordinary tool activity just because segment ids changed", () => {
+    const messages: UIMessage[] = [
+      {
+        id: "r1",
+        role: "assistant",
+        content: "",
+        reasoning: "first pass",
+        activitySegmentId: "seg-1",
+        createdAt: 1,
+      },
+      {
+        id: "t1",
+        role: "tool",
+        kind: "trace",
+        content: "read_file()",
+        traces: ["read_file()"],
+        activitySegmentId: "seg-1",
+        createdAt: 2,
+      },
+      {
+        id: "r2",
+        role: "assistant",
+        content: "",
+        reasoning: "second pass",
+        activitySegmentId: "seg-2",
+        createdAt: 3,
+      },
+      {
+        id: "t2",
+        role: "tool",
+        kind: "trace",
+        content: "grep()",
+        traces: ["grep()"],
+        activitySegmentId: "seg-2",
+        createdAt: 4,
+      },
+    ];
+
+    const units = buildDisplayUnits(messages);
+
+    expect(units).toHaveLength(1);
+    expect(units[0].type === "cluster" ? units[0].messages.map((m) => m.id) : []).toEqual([
+      "r1",
+      "t1",
+      "r2",
+      "t2",
+    ]);
+  });
+
+  it("only marks the current activity cluster as live while streaming", () => {
+    const messages: UIMessage[] = [
+      {
+        id: "r1",
+        role: "assistant",
+        content: "",
+        reasoning: "first pass",
+        reasoningStreaming: true,
+        activitySegmentId: "seg-1",
+        createdAt: 1,
+      },
+      {
+        id: "t1",
+        role: "tool",
+        kind: "trace",
+        content: "edit_file()",
+        traces: ["edit_file()"],
+        fileEdits: [{
+          call_id: "call-edit",
+          tool: "edit_file",
+          path: "foo.txt",
+          phase: "start",
+          added: 4,
+          deleted: 1,
+          approximate: true,
+          status: "editing",
+        }],
+        activitySegmentId: "seg-1",
+        createdAt: 2,
+      },
+      {
+        id: "r2",
+        role: "assistant",
+        content: "",
+        reasoning: "second pass",
+        reasoningStreaming: true,
+        activitySegmentId: "seg-2",
+        createdAt: 3,
+      },
+    ];
+
+    render(<ThreadMessages messages={messages} isStreaming />);
+
+    expect(screen.getByRole("button", { name: /edited foo\.txt/i })).toBeInTheDocument();
+    expect(screen.queryByRole("button", { name: /editing foo\.txt/i })).not.toBeInTheDocument();
+    expect(screen.getByRole("button", { name: /working/i })).toBeInTheDocument();
+  });
+
+  it("folds final answer reasoning into the preceding activity cluster", () => {
+    const messages: UIMessage[] = [
+      {
+        id: "r1",
+        role: "assistant",
+        content: "",
+        reasoning: "search plan",
+        reasoningStreaming: false,
+        createdAt: 1,
+      },
+      {
+        id: "t1",
+        role: "tool",
+        kind: "trace",
+        content: "web_search()",
+        traces: ["web_search()"],
+        createdAt: 2,
+      },
+      {
+        id: "a1",
+        role: "assistant",
+        content: "final answer",
+        reasoning: "summarize results",
+        reasoningStreaming: false,
+        createdAt: 3,
+      },
+    ];
+
+    const units = buildDisplayUnits(messages);
+
+    expect(units).toHaveLength(2);
+    expect(units[0]).toMatchObject({ type: "cluster" });
+    expect(units[0].type === "cluster" ? units[0].messages.map((m) => m.id) : []).toEqual([
+      "r1",
+      "t1",
+      "a1-reasoning",
+    ]);
+    expect(units[1]).toMatchObject({
+      type: "single",
+      message: {
+        id: "a1",
+        content: "final answer",
+      },
+    });
+    if (units[1].type === "single") {
+      expect(units[1].message).not.toHaveProperty("reasoning");
+    }
+
+    render(<ThreadMessages messages={messages} isStreaming={false} />);
+    expect(screen.queryByRole("button", { name: /^thinking$/i })).not.toBeInTheDocument();
+    expect(screen.getByText("final answer")).toBeInTheDocument();
+  });
+
   it("shows copy only on the last assistant slice before the next user turn", () => {
     const messages: UIMessage[] = [
       {
@@ -89,4 +293,37 @@ describe("ThreadMessages", () => {
     render(<ThreadMessages messages={messages} isStreaming={false} />);
     expect(screen.getAllByRole("button", { name: "Copy reply" })).toHaveLength(1);
   });
+
+  it("computes final assistant copy flags with user-boundary semantics", () => {
+    const units = buildDisplayUnits([
+      { id: "u1", role: "user", content: "one", createdAt: 1 },
+      { id: "a1", role: "assistant", content: "draft", createdAt: 2 },
+      {
+        id: "t1",
+        role: "tool",
+        kind: "trace",
+        content: "tool()",
+        traces: ["tool()"],
+        createdAt: 3,
+      },
+      { id: "a2", role: "assistant", content: "final", createdAt: 4 },
+      { id: "u2", role: "user", content: "two", createdAt: 5 },
+      { id: "a3", role: "assistant", content: "next", createdAt: 6 },
+    ]);
+
+    const flags = assistantCopyFlags(units);
+    const assistantFlags = units
+      .map((unit, index) =>
+        unit.type === "single" && unit.message.role === "assistant"
+          ? [unit.message.id, flags[index]]
+          : null,
+      )
+      .filter(Boolean);
+
+    expect(assistantFlags).toEqual([
+      ["a1", false],
+      ["a2", true],
+      ["a3", true],
+    ]);
+  });
 });

webui/src/tests/thread-shell.test.tsx

@@ -8,7 +8,7 @@ import type { UIMessage } from "@/lib/types";
 function makeClient() {
   const errorHandlers = new Set<(err: { kind: string }) => void>();
   const chatHandlers = new Map<string, Set<(ev: import("@/lib/types").InboundEvent) => void>>();
-  const sessionUpdateHandlers = new Set<(chatId: string) => void>();
+  const sessionUpdateHandlers = new Set<(chatId: string, scope?: string) => void>();
   const goalStateByChatId = new Map<string, import("@/lib/types").GoalStateWsPayload>();
   return {
     status: "open" as const,
@@ -34,7 +34,7 @@ function makeClient() {
         errorHandlers.delete(handler);
       };
     },
-    onSessionUpdate: (handler: (chatId: string) => void) => {
+    onSessionUpdate: (handler: (chatId: string, scope?: string) => void) => {
       sessionUpdateHandlers.add(handler);
       return () => {
         sessionUpdateHandlers.delete(handler);
@@ -49,8 +49,8 @@ function makeClient() {
       }
       for (const h of chatHandlers.get(chatId) ?? []) h(ev);
     },
-    _emitSessionUpdate(chatId: string) {
-      for (const h of sessionUpdateHandlers) h(chatId);
+    _emitSessionUpdate(chatId: string, scope?: string) {
+      for (const h of sessionUpdateHandlers) h(chatId, scope);
     },
     sendMessage: vi.fn(),
     newChat: vi.fn(),
@@ -651,6 +651,52 @@ describe("ThreadShell", () => {
     expect(historyCalls).toBe(1);
   });
 
+  it("does not refetch thread history for metadata-only session updates", async () => {
+    const client = makeClient();
+    let historyCalls = 0;
+    vi.stubGlobal(
+      "fetch",
+      vi.fn(async (input: RequestInfo | URL) => {
+        const url = String(input);
+        if (url.includes("websocket%3Achat-a/webui-thread")) {
+          historyCalls += 1;
+          return httpJson(
+            transcriptFromSimpleMessages([
+              { role: "user", content: "question" },
+              { role: "assistant", content: "answer" },
+            ]),
+          );
+        }
+        return {
+          ok: false,
+          status: 404,
+          json: async () => ({}),
+        };
+      }),
+    );
+
+    render(
+      wrap(
+        client,
+        <ThreadShell
+          session={session("chat-a")}
+          title="Chat chat-a"
+          onToggleSidebar={() => {}}
+          onNewChat={() => {}}
+        />,
+      ),
+    );
+
+    await waitFor(() => expect(screen.getByText("answer")).toBeInTheDocument());
+    expect(historyCalls).toBe(1);
+
+    await act(async () => {
+      client._emitSessionUpdate("chat-a", "metadata");
+    });
+
+    expect(historyCalls).toBe(1);
+  });
+
   it("scrolls to the bottom after loading a session from the blank new-chat page", async () => {
     const client = makeClient();
     const scrollIntoView = vi.fn();

webui/src/tests/thread-viewport.test.tsx

@@ -1,7 +1,12 @@
-import { act, render, waitFor } from "@testing-library/react";
+import { act, fireEvent, render, screen, waitFor } from "@testing-library/react";
 import { describe, expect, it, vi } from "vitest";
 
-import { ThreadViewport } from "@/components/thread/ThreadViewport";
+import {
+  HISTORY_WINDOW_INCREMENT,
+  INITIAL_HISTORY_WINDOW,
+  ThreadViewport,
+  windowMessages,
+} from "@/components/thread/ThreadViewport";
 import type { UIMessage } from "@/lib/types";
 
 const messages: UIMessage[] = [
@@ -15,7 +20,191 @@ const messages: UIMessage[] = [
 
 const emptyMessages: UIMessage[] = [];
 
+interface ResizeObserverInstance {
+  element?: Element;
+  callback: ResizeObserverCallback;
+  disconnect: ReturnType<typeof vi.fn>;
+}
+
+function makeLongMessages(count: number): UIMessage[] {
+  return Array.from({ length: count }, (_, index) => ({
+    id: `m${index}`,
+    role: "user" as const,
+    content: `message ${index}`,
+    createdAt: index,
+  }));
+}
+
 describe("ThreadViewport", () => {
+  it("keeps the scroll-to-bottom button above a growing composer", () => {
+    const originalResizeObserver = globalThis.ResizeObserver;
+    const resizeObservers: ResizeObserverInstance[] = [];
+    class MockResizeObserver {
+      element?: Element;
+      callback: ResizeObserverCallback;
+      disconnect = vi.fn();
+
+      constructor(callback: ResizeObserverCallback) {
+        this.callback = callback;
+        resizeObservers.push(this);
+      }
+
+      observe(element: Element) {
+        this.element = element;
+      }
+    }
+    vi.stubGlobal("ResizeObserver", MockResizeObserver);
+
+    try {
+      const { container } = render(
+        <ThreadViewport
+          messages={messages}
+          isStreaming={false}
+          composer={<div>composer</div>}
+        />,
+      );
+      const scroller = container.firstElementChild?.firstElementChild as HTMLElement;
+      Object.defineProperties(scroller, {
+        scrollHeight: { configurable: true, value: 2400 },
+        clientHeight: { configurable: true, value: 600 },
+        scrollTop: { configurable: true, value: 0 },
+      });
+
+      act(() => {
+        scroller.dispatchEvent(new Event("scroll"));
+      });
+
+      const button = screen.getByRole("button", { name: "Scroll to bottom" });
+      expect(button).toHaveStyle({ bottom: "192px" });
+
+      const composerDock = screen.getByTestId("thread-composer-dock");
+      composerDock.getBoundingClientRect = () =>
+        ({
+          height: 240,
+          width: 800,
+          top: 0,
+          right: 800,
+          bottom: 240,
+          left: 0,
+          x: 0,
+          y: 0,
+          toJSON: () => ({}),
+        }) as DOMRect;
+
+      const composerObserver = resizeObservers.find(
+        (observer) => observer.element === composerDock,
+      );
+      expect(composerObserver).toBeDefined();
+
+      act(() => {
+        composerObserver!.callback([], composerObserver as unknown as ResizeObserver);
+      });
+
+      expect(button).toHaveStyle({ bottom: "256px" });
+    } finally {
+      vi.stubGlobal("ResizeObserver", originalResizeObserver);
+    }
+  });
+
+  it("hides the scroll-to-bottom button when disabled for the welcome view", () => {
+    const { container } = render(
+      <ThreadViewport
+        messages={emptyMessages}
+        isStreaming={false}
+        composer={<div>composer</div>}
+        emptyState={<div>welcome</div>}
+        showScrollToBottomButton={false}
+      />,
+    );
+    const scroller = container.firstElementChild?.firstElementChild as HTMLElement;
+    Object.defineProperties(scroller, {
+      scrollHeight: { configurable: true, value: 2400 },
+      clientHeight: { configurable: true, value: 600 },
+      scrollTop: { configurable: true, value: 0 },
+    });
+
+    act(() => {
+      scroller.dispatchEvent(new Event("scroll"));
+    });
+
+    expect(screen.queryByRole("button", { name: "Scroll to bottom" })).not.toBeInTheDocument();
+  });
+
+  it("renders only the tail window for long history by default", () => {
+    const longMessages = makeLongMessages(300);
+
+    render(
+      <ThreadViewport
+        messages={longMessages}
+        isStreaming={false}
+        composer={<div />}
+      />,
+    );
+
+    expect(screen.queryByText("message 139")).not.toBeInTheDocument();
+    expect(screen.getByText("message 140")).toBeInTheDocument();
+    expect(screen.getByText("message 299")).toBeInTheDocument();
+    expect(screen.getByRole("button", { name: "Load earlier messages" })).toBeInTheDocument();
+  });
+
+  it("loads earlier history in fixed increments without rendering the whole transcript", () => {
+    const longMessages = makeLongMessages(300);
+
+    render(
+      <ThreadViewport
+        messages={longMessages}
+        isStreaming={false}
+        composer={<div />}
+      />,
+    );
+
+    fireEvent.click(screen.getByRole("button", { name: "Load earlier messages" }));
+
+    const firstVisible =
+      300 - INITIAL_HISTORY_WINDOW - HISTORY_WINDOW_INCREMENT;
+
+    expect(
+      screen.queryByText(`message ${firstVisible - 1}`),
+    ).not.toBeInTheDocument();
+    expect(screen.getByText(`message ${firstVisible}`)).toBeInTheDocument();
+    expect(screen.getByText("message 299")).toBeInTheDocument();
+  });
+
+  it("expands the window start to avoid cutting an agent activity cluster", () => {
+    const clustered = makeLongMessages(200);
+    clustered.splice(
+      38,
+      3,
+      {
+        id: "r0",
+        role: "assistant",
+        content: "",
+        reasoning: "first reasoning",
+        createdAt: 38,
+      },
+      {
+        id: "t0",
+        role: "tool",
+        kind: "trace",
+        content: "tool()",
+        traces: ["tool()"],
+        createdAt: 39,
+      },
+      {
+        id: "r1",
+        role: "assistant",
+        content: "",
+        reasoning: "second reasoning",
+        createdAt: 40,
+      },
+    );
+
+    const visible = windowMessages(clustered, INITIAL_HISTORY_WINDOW);
+
+    expect(visible[0].id).toBe("r0");
+    expect(visible).toHaveLength(INITIAL_HISTORY_WINDOW + 2);
+  });
+
   it("resets to the bottom when opening a different conversation", async () => {
     const scrollIntoView = vi.fn();
     const originalScrollIntoView = HTMLElement.prototype.scrollIntoView;

webui/src/tests/useDeferredTitleRefresh.test.tsx

@@ -0,0 +1,110 @@
+import { act, renderHook } from "@testing-library/react";
+import { afterEach, beforeEach, describe, expect, it, vi } from "vitest";
+
+import { useDeferredTitleRefresh } from "@/hooks/useDeferredTitleRefresh";
+import type { ChatSummary } from "@/lib/types";
+
+function session(overrides: Partial<ChatSummary> = {}): ChatSummary {
+  return {
+    key: "websocket:chat-a",
+    channel: "websocket",
+    chatId: "chat-a",
+    createdAt: null,
+    updatedAt: null,
+    title: "",
+    preview: "First user message",
+    ...overrides,
+  };
+}
+
+describe("useDeferredTitleRefresh", () => {
+  beforeEach(() => {
+    vi.useFakeTimers();
+  });
+
+  afterEach(() => {
+    vi.useRealTimers();
+  });
+
+  it("retries refreshing untitled sessions after turn_end", () => {
+    const refresh = vi.fn().mockResolvedValue(undefined);
+    const { result } = renderHook(() =>
+      useDeferredTitleRefresh(session(), refresh, [100, 300]),
+    );
+
+    act(() => {
+      result.current();
+    });
+
+    expect(refresh).toHaveBeenCalledTimes(1);
+
+    act(() => {
+      vi.advanceTimersByTime(100);
+    });
+    expect(refresh).toHaveBeenCalledTimes(2);
+
+    act(() => {
+      vi.advanceTimersByTime(200);
+    });
+    expect(refresh).toHaveBeenCalledTimes(3);
+  });
+
+  it("stops pending retries once a generated title arrives", () => {
+    const refresh = vi.fn().mockResolvedValue(undefined);
+    const { result, rerender } = renderHook(
+      ({ activeSession }) =>
+        useDeferredTitleRefresh(activeSession, refresh, [100, 300]),
+      { initialProps: { activeSession: session() } },
+    );
+
+    act(() => {
+      result.current();
+    });
+    rerender({ activeSession: session({ title: "Generated title" }) });
+
+    act(() => {
+      vi.advanceTimersByTime(300);
+    });
+
+    expect(refresh).toHaveBeenCalledTimes(1);
+  });
+
+  it("does not retry when the active session already has a title", () => {
+    const refresh = vi.fn().mockResolvedValue(undefined);
+    const { result } = renderHook(() =>
+      useDeferredTitleRefresh(session({ title: "Existing title" }), refresh, [100]),
+    );
+
+    act(() => {
+      result.current();
+      vi.advanceTimersByTime(100);
+    });
+
+    expect(refresh).toHaveBeenCalledTimes(1);
+  });
+
+  it("clears pending retries when the active chat changes", () => {
+    const refresh = vi.fn().mockResolvedValue(undefined);
+    const { result, rerender } = renderHook(
+      ({ activeSession }) =>
+        useDeferredTitleRefresh(activeSession, refresh, [100]),
+      { initialProps: { activeSession: session() } },
+    );
+
+    act(() => {
+      result.current();
+    });
+    rerender({
+      activeSession: session({
+        key: "websocket:chat-b",
+        chatId: "chat-b",
+      }),
+    });
+
+    act(() => {
+      vi.advanceTimersByTime(100);
+    });
+
+    expect(refresh).toHaveBeenCalledTimes(1);
+  });
+});

webui/src/tests/useNanobotStream.test.tsx

@@ -83,7 +83,112 @@ function wrap(client: ReturnType<typeof fakeClient>["client"]) {
   };
 }
 
+async function flushStreamFrame() {
+  await act(async () => {
+    await new Promise<void>((resolve) => {
+      requestAnimationFrame(() => resolve());
+    });
+  });
+}
+
 describe("useNanobotStream", () => {
+  it("batches answer deltas into one animation-frame update", async () => {
+    const fake = fakeClient();
+    const requestFrame = vi.spyOn(window, "requestAnimationFrame");
+    const { result } = renderHook(() => useNanobotStream("chat-batch", EMPTY_MESSAGES), {
+      wrapper: wrap(fake.client),
+    });
+
+    act(() => {
+      fake.emit("chat-batch", {
+        event: "delta",
+        chat_id: "chat-batch",
+        text: "Hello",
+      });
+      fake.emit("chat-batch", {
+        event: "delta",
+        chat_id: "chat-batch",
+        text: " world",
+      });
+    });
+
+    expect(requestFrame).toHaveBeenCalledTimes(1);
+    expect(result.current.messages).toHaveLength(0);
+
+    await flushStreamFrame();
+
+    expect(result.current.messages).toHaveLength(1);
+    expect(result.current.messages[0]).toMatchObject({
+      role: "assistant",
+      content: "Hello world",
+      isStreaming: true,
+    });
+    requestFrame.mockRestore();
+  });
+
+  it("flushes pending delta text before turn_end finalizes the turn", () => {
+    const fake = fakeClient();
+    const { result } = renderHook(() => useNanobotStream("chat-flush", EMPTY_MESSAGES), {
+      wrapper: wrap(fake.client),
+    });
+
+    act(() => {
+      fake.emit("chat-flush", {
+        event: "delta",
+        chat_id: "chat-flush",
+        text: "final chunk",
+      });
+      fake.emit("chat-flush", {
+        event: "turn_end",
+        chat_id: "chat-flush",
+      });
+    });
+
+    expect(result.current.messages).toHaveLength(1);
+    expect(result.current.messages[0]).toMatchObject({
+      role: "assistant",
+      content: "final chunk",
+      isStreaming: false,
+    });
+    expect(result.current.isStreaming).toBe(false);
+  });
+
+  it("drops pending stream work when switching chats", async () => {
+    const fake = fakeClient();
+    const { result, rerender } = renderHook(
+      ({ chatId }: { chatId: string }) => useNanobotStream(chatId, EMPTY_MESSAGES),
+      {
+        wrapper: wrap(fake.client),
+        initialProps: { chatId: "chat-old" },
+      },
+    );
+
+    act(() => {
+      fake.emit("chat-old", {
+        event: "delta",
+        chat_id: "chat-old",
+        text: "stale",
+      });
+    });
+
+    rerender({ chatId: "chat-new" });
+
+    act(() => {
+      fake.emit("chat-new", {
+        event: "delta",
+        chat_id: "chat-new",
+        text: "fresh",
+      });
+    });
+    await flushStreamFrame();
+
+    expect(result.current.messages).toHaveLength(1);
+    expect(result.current.messages[0]).toMatchObject({
+      role: "assistant",
+      content: "fresh",
+    });
+  });
+
   it("starts in streaming mode when history shows pending tool calls", () => {
     const fake = fakeClient();
     const initialMessages = [{
@@ -203,7 +308,174 @@ describe("useNanobotStream", () => {
     );
   });
 
-  it("accumulates reasoning_delta chunks on a placeholder until reasoning_end", () => {
+  it("renders live file_edit events as their own activity trace", () => {
+    const fake = fakeClient();
+    const { result } = renderHook(() => useNanobotStream("chat-file-edit", EMPTY_MESSAGES), {
+      wrapper: wrap(fake.client),
+    });
+
+    act(() => {
+      fake.emit("chat-file-edit", {
+        event: "message",
+        chat_id: "chat-file-edit",
+        text: 'write_file({"path":"foo.txt"})',
+        kind: "tool_hint",
+      });
+      fake.emit("chat-file-edit", {
+        event: "file_edit",
+        chat_id: "chat-file-edit",
+        edits: [{
+          call_id: "call-write",
+          tool: "write_file",
+          path: "foo.txt",
+          phase: "start",
+          added: 1,
+          deleted: 0,
+          approximate: true,
+          status: "editing",
+        }],
+      });
+      fake.emit("chat-file-edit", {
+        event: "file_edit",
+        chat_id: "chat-file-edit",
+        edits: [{
+          call_id: "call-write",
+          tool: "write_file",
+          path: "foo.txt",
+          phase: "end",
+          added: 3,
+          deleted: 1,
+          approximate: false,
+          status: "done",
+        }],
+      });
+    });
+
+    expect(result.current.messages).toHaveLength(2);
+    expect(result.current.messages[0]).toMatchObject({
+      role: "tool",
+      kind: "trace",
+      traces: ['write_file({"path":"foo.txt"})'],
+    });
+    expect(result.current.messages[1]).toMatchObject({
+      role: "tool",
+      kind: "trace",
+      fileEdits: [{
+        call_id: "call-write",
+        status: "done",
+        added: 3,
+        deleted: 1,
+        approximate: false,
+      }],
+    });
+    expect(result.current.messages[1].activitySegmentId).toBeTruthy();
+    expect(result.current.messages[1].activitySegmentId).not.toBe(
+      result.current.messages[0].activitySegmentId,
+    );
+  });
+
+  it("starts a new assistant bubble for deltas after stream_end and activity", async () => {
+    const fake = fakeClient();
+    const { result } = renderHook(() => useNanobotStream("chat-stream-segments", EMPTY_MESSAGES), {
+      wrapper: wrap(fake.client),
+    });
+
+    act(() => {
+      fake.emit("chat-stream-segments", {
+        event: "delta",
+        chat_id: "chat-stream-segments",
+        text: "I created the files.",
+      });
+      fake.emit("chat-stream-segments", {
+        event: "stream_end",
+        chat_id: "chat-stream-segments",
+      });
+      fake.emit("chat-stream-segments", {
+        event: "message",
+        chat_id: "chat-stream-segments",
+        text: 'write_file({"path":"minecraft-fps/options.txt"})',
+        kind: "tool_hint",
+      });
+      fake.emit("chat-stream-segments", {
+        event: "delta",
+        chat_id: "chat-stream-segments",
+        text: "Now I will summarize the edits.",
+      });
+    });
+
+    await flushStreamFrame();
+
+    expect(result.current.messages).toHaveLength(3);
+    expect(result.current.messages[0]).toMatchObject({
+      role: "assistant",
+      content: "I created the files.",
+    });
+    expect(result.current.messages[1]).toMatchObject({
+      role: "tool",
+      kind: "trace",
+      traces: ['write_file({"path":"minecraft-fps/options.txt"})'],
+    });
+    expect(result.current.messages[2]).toMatchObject({
+      role: "assistant",
+      content: "Now I will summarize the edits.",
+    });
+  });
+
+  it("opens a new activity segment for reasoning after file edit activity", async () => {
+    const fake = fakeClient();
+    const { result } = renderHook(() => useNanobotStream("chat-file-segments", EMPTY_MESSAGES), {
+      wrapper: wrap(fake.client),
+    });
+
+    act(() => {
+      fake.emit("chat-file-segments", {
+        event: "reasoning_delta",
+        chat_id: "chat-file-segments",
+        text: "Plan.",
+      });
+      fake.emit("chat-file-segments", {
+        event: "reasoning_end",
+        chat_id: "chat-file-segments",
+      });
+      fake.emit("chat-file-segments", {
+        event: "message",
+        chat_id: "chat-file-segments",
+        text: 'edit_file({"path":"foo.txt"})',
+        kind: "tool_hint",
+      });
+      fake.emit("chat-file-segments", {
+        event: "file_edit",
+        chat_id: "chat-file-segments",
+        edits: [{
+          call_id: "call-edit",
+          tool: "edit_file",
+          path: "foo.txt",
+          phase: "start",
+          added: 1,
+          deleted: 1,
+          approximate: true,
+          status: "editing",
+        }],
+      });
+      fake.emit("chat-file-segments", {
+        event: "reasoning_delta",
+        chat_id: "chat-file-segments",
+        text: "Review result.",
+      });
+    });
+
+    await flushStreamFrame();
+
+    expect(result.current.messages).toHaveLength(4);
+    const firstSegment = result.current.messages[0].activitySegmentId;
+    expect(firstSegment).toBeTruthy();
+    expect(result.current.messages[1].activitySegmentId).toBe(firstSegment);
+    expect(result.current.messages[2].activitySegmentId).toBeTruthy();
+    expect(result.current.messages[2].activitySegmentId).not.toBe(firstSegment);
+    expect(result.current.messages[3].activitySegmentId).toBe(firstSegment);
+  });
+
+  it("accumulates reasoning_delta chunks on a placeholder until reasoning_end", async () => {
     const fake = fakeClient();
     const { result } = renderHook(() => useNanobotStream("chat-r", EMPTY_MESSAGES), {
       wrapper: wrap(fake.client),
@@ -222,6 +494,8 @@ describe("useNanobotStream", () => {
       });
     });
 
+    await flushStreamFrame();
+
     expect(result.current.messages).toHaveLength(1);
     expect(result.current.messages[0].role).toBe("assistant");
     expect(result.current.messages[0].reasoning).toBe("Let me think step by step.");
@@ -328,7 +602,7 @@ describe("useNanobotStream", () => {
     expect(result.current.messages[0].reasoningStreaming).toBe(false);
   });
 
-  it("does not attach a new turn's reasoning across the latest user boundary", () => {
+  it("does not attach a new turn's reasoning across the latest user boundary", async () => {
     const fake = fakeClient();
     const initialMessages = [
       {
@@ -358,6 +632,8 @@ describe("useNanobotStream", () => {
       });
     });
 
+    await flushStreamFrame();
+
     expect(result.current.messages).toHaveLength(3);
     expect(result.current.messages[0].reasoning).toBe("Previous thought.");
     expect(result.current.messages[2].role).toBe("assistant");
@@ -366,7 +642,7 @@ describe("useNanobotStream", () => {
     expect(result.current.messages[2].reasoningStreaming).toBe(true);
   });
 
-  it("does not attach reasoning across a tool trace boundary", () => {
+  it("does not attach reasoning across a tool trace boundary", async () => {
     const fake = fakeClient();
     const { result } = renderHook(() => useNanobotStream("chat-r7", EMPTY_MESSAGES), {
       wrapper: wrap(fake.client),
@@ -392,6 +668,8 @@ describe("useNanobotStream", () => {
       });
     });
 
+    await flushStreamFrame();
+
     expect(result.current.messages).toHaveLength(3);
     expect(result.current.messages.map((m) => m.kind ?? "message")).toEqual([
       "message",
@@ -651,7 +929,7 @@ describe("useNanobotStream", () => {
     expect(result.current.messages[0].content).toBe("long task");
   });
 
-  it("keeps streaming alive across stream_end and completes on turn_end", () => {
+  it("keeps streaming alive across stream_end and completes on turn_end", async () => {
     const fake = fakeClient();
     const onTurnEnd = vi.fn();
     const { result } = renderHook(() => useNanobotStream("chat-s", EMPTY_MESSAGES, false, onTurnEnd), {
@@ -666,6 +944,8 @@ describe("useNanobotStream", () => {
       });
     });
 
+    await flushStreamFrame();
+
     expect(result.current.isStreaming).toBe(true);
     expect(result.current.messages[0]).toMatchObject({
       role: "assistant",

webui/src/tests/useSessions.test.tsx

@@ -2,7 +2,7 @@ import { act, renderHook, waitFor } from "@testing-library/react";
 import type { ReactNode } from "react";
 import { beforeEach, describe, expect, it, vi } from "vitest";
 
-import { useSessionHistory, useSessions } from "@/hooks/useSessions";
+import { sessionTitle, useSessionHistory, useSessions } from "@/hooks/useSessions";
 import * as api from "@/lib/api";
 import { ClientProvider } from "@/providers/ClientProvider";
 
@@ -17,7 +17,7 @@ vi.mock("@/lib/api", async (importOriginal) => {
 });
 
 function fakeClient() {
-  const sessionUpdateHandlers = new Set<(chatId: string) => void>();
+  const sessionUpdateHandlers = new Set<(chatId: string, scope?: string) => void>();
   return {
     status: "open" as const,
     defaultChatId: null as string | null,
@@ -25,12 +25,12 @@ function fakeClient() {
     onError: () => () => {},
     onChat: () => () => {},
     getRunStartedAt: () => null,
-    onSessionUpdate: (handler: (chatId: string) => void) => {
+    onSessionUpdate: (handler: (chatId: string, scope?: string) => void) => {
       sessionUpdateHandlers.add(handler);
       return () => sessionUpdateHandlers.delete(handler);
     },
-    emitSessionUpdate: (chatId: string) => {
-      for (const handler of sessionUpdateHandlers) handler(chatId);
+    emitSessionUpdate: (chatId: string, scope?: string) => {
+      for (const handler of sessionUpdateHandlers) handler(chatId, scope);
     },
     sendMessage: vi.fn(),
     newChat: vi.fn(),
@@ -61,6 +61,28 @@ describe("useSessions", () => {
     vi.mocked(api.fetchWebuiThread).mockReset();
   });
 
+  it("does not use low-information greetings as fallback session titles", () => {
+    expect(sessionTitle({
+      key: "websocket:chat-hi",
+      channel: "websocket",
+      chatId: "chat-hi",
+      createdAt: "2026-04-16T10:00:00Z",
+      updatedAt: "2026-04-16T10:00:00Z",
+      title: "",
+      preview: "hi",
+    })).toBe("New chat");
+
+    expect(sessionTitle({
+      key: "websocket:chat-work",
+      channel: "websocket",
+      chatId: "chat-work",
+      createdAt: "2026-04-16T10:00:00Z",
+      updatedAt: "2026-04-16T10:00:00Z",
+      title: "",
+      preview: "帮我优化 WebUI 性能",
+    })).toBe("帮我优化 WebUI 性能");
+  });
+
   it("removes a session from the local list after delete succeeds", async () => {
     vi.mocked(api.listSessions).mockResolvedValue([
       {

webui/src/types/react-syntax-highlighter-subpaths.d.ts

@@ -0,0 +1,22 @@
+declare module "react-syntax-highlighter/dist/esm/prism-async-light" {
+  import * as React from "react";
+  import type { SyntaxHighlighterProps } from "react-syntax-highlighter";
+
+  export default class SyntaxHighlighter extends React.Component<SyntaxHighlighterProps> {
+    static registerLanguage(name: string, func: unknown): void;
+  }
+}
+
+declare module "react-syntax-highlighter/dist/esm/styles/prism/one-dark" {
+  import type * as React from "react";
+
+  const style: { [key: string]: React.CSSProperties };
+  export default style;
+}
+
+declare module "react-syntax-highlighter/dist/esm/styles/prism/one-light" {
+  import type * as React from "react";
+
+  const style: { [key: string]: React.CSSProperties };
+  export default style;
+}

webui/tsconfig.build.json

@@ -1,7 +1,7 @@
 {
   "extends": "./tsconfig.json",
   "compilerOptions": {
-    "types": ["node"]
+    "types": ["node", "vite/client"]
   },
   "exclude": ["src/tests/**"]
 }

webui/vite.config.ts

@@ -25,6 +25,36 @@ export default defineConfig(({ mode }) => {
       outDir: path.resolve(__dirname, "../nanobot/web/dist"),
       emptyOutDir: true,
       sourcemap: false,
+      rollupOptions: {
+        output: {
+          manualChunks(id) {
+            if (id.includes("node_modules/refractor/lang/")) {
+              return;
+            }
+            if (
+              id.includes("node_modules/react-syntax-highlighter")
+              || id.includes("node_modules/refractor/core")
+            ) {
+              return "syntax-highlight";
+            }
+            if (
+              id.includes("node_modules/react-markdown")
+              || id.includes("node_modules/remark-")
+              || id.includes("node_modules/rehype-")
+              || id.includes("node_modules/unified")
+              || id.includes("node_modules/mdast-")
+              || id.includes("node_modules/hast-")
+              || id.includes("node_modules/micromark")
+              || id.includes("node_modules/unist-")
+            ) {
+              return "markdown-vendor";
+            }
+            if (id.includes("node_modules/katex")) {
+              return "katex";
+            }
+          },
+        },
+      },
     },
     server: {
       host: "127.0.0.1",