When enabled, all /_web/* routes require Bearer token authentication,
same as /api/v1/*. This hardens security for server/headless/public
deployments where the same-origin assumption doesn't hold.
- Add `require_auth` field to config schema (default: false)
- Add `setRequireAuth()` to auth hook, conditionally enforce auth on /_web/*
- Support `--require-auth` in both `chatlab start` and `--daemon` mode
- Read from config.api.require_auth or CLI flag override
Co-authored-by: Cursor <cursoragent@cursor.com>
在 Web 生产模式（chatlab start）下，浏览器无法直连 chatlab.fun（CORS 限制），
导致首页底部社交链接（如小红书）及云端助手/技能预设拉取失败。
Vite dev proxy 已有对应配置，但生产 Fastify server 缺少该代理，现补全。
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
When the user selects 'auto' thinking level for an effort-based reasoning
model (Kimi, Doubao) that has thinkingLevelMap but no thinkingFormat,
the previous code fell through to `return undefined`, causing pi-agent-core
to treat it as 'off' and send reasoning_effort:'none' to the API —
silently disabling thinking.
pi-ai's clampThinkingLevel only covers EXTENDED_THINKING_LEVELS (no 'auto'),
so 'auto' cannot be forwarded verbatim. Return 'high' instead, matching
the behavior for thinkingFormat models, to at least enable reasoning.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Custom models auto-saved by ensureCustomProvidersAndModels carry only
capabilities:['chat'], so inferReasoning was returning reasoning=false
even for models like qwen3:8b whose names match the reasoning heuristic.
The UI still showed thinking levels via getSupportedThinkingLevels, but
selections were silently ignored by the backend.
Fix: for non-builtin (builtin=false) model definitions, also apply the
name/provider heuristic as a fallback alongside the capabilities check.
Builtin catalog models remain authoritative.
Adds a regression test covering the exact findModelFn shape produced by
ensureCustomProvidersAndModels.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The supportedThinkingLevels, modelContextWindow, and model name display
were reading from config.model (static config value) instead of the
slot's modelId (actual user-selected model). This caused the thinking
selector to show/hide incorrectly when users switched models.
Also fix HTTP utils sending Content-Type: application/json with empty
body when body is null (e.g. setFastModel(null)), which caused Fastify
to return 400.
Co-authored-by: Cursor <cursoragent@cursor.com>
- StatusBar model button: show model name instead of provider, click opens
settings directly (scrolls to default model section), remove dropdown
- SessionIndexSection: replace N serial pluginQuery calls with single batch
API (GET /_web/sessions/index-stats), add IntersectionObserver lazy loading
- Full-stack implementation: CLI route, Electron IPC/Worker, frontend adapters
Co-authored-by: Cursor <cursoragent@cursor.com>
- Add 'default' option (no reasoning params sent, model uses native behavior)
- Add 'auto' option for models that natively support it (deepseek_hybrid, kimi, doubao)
- Expand model recognition with regex coverage for 15+ model families
- Add Gemini 2.5+ thinking level support via native Google provider
- Simplify deepseek_hybrid options to ['default', 'off', 'auto'] aligned with Cherry Studio
- Use primary theme color for thinking selector instead of hardcoded blue
- Update docs with thinking control architecture section
Co-authored-by: Cursor <cursoragent@cursor.com>
Adds a lightbulb button next to the model switcher in ChatStatusBar
that lets users choose the thinking intensity for each reasoning model,
with the choice remembered per model slot (configId:modelId).
Key changes:
- packages/core/src/ai/thinking.ts (new): per-model-type thinking level
table inspired by cherry-studio's MODEL_SUPPORTED_REASONING_EFFORT.
Covers Qwen/GLM (on/off), DeepSeek V4 (high/max), o-series
(low/medium/high), gpt-5.x (minimal–xhigh), Kimi/Doubao/Grok and a
default bucket. Exports getSupportedThinkingLevels, isReasoningModel,
getThinkingCompat.
- packages/node-runtime/src/ai/llm-builder.ts: inferReasoning now calls
getThinkingCompat() which sets compat.supportsReasoningEffort and
compat.thinkingLevelMap so pi-ai can inject the correct request param
(reasoning_effort for OpenAI/DeepSeek, enable_thinking for Qwen/GLM).
- packages/node-runtime/src/ai/agent/types.ts + core.ts: AgentCoreOptions
gains an optional thinkingLevel; core.ts:93 uses clampThinkingLevel()
(from pi-ai) to honour it safely.
- Full transparent pass-through: desktop Agent → IPC handler → preload
→ agentApi.runStream; CLI HTTP route → runServerAgent → runAgentCore.
- src/stores/prompt.ts: thinkingLevels Record<configId:modelId,
ThinkingLevel> persisted in backendPersist; getThinkingLevel/
setThinkingLevel helpers.
- src/stores/aiChat.ts: reads current model slot and passes remembered
level to both sendMessage and editCurrentRoundOnly runStream calls.
- ChatStatusBar.vue: UPopover thinking selector appears only for
reasoning models, shows only the levels that model supports, highlights
current choice, hides for non-reasoning models.
- 46 tests passing (26 new in thinking.test.ts + 20 in llm-builder).
Scope: openai-completions path only (Qwen, DeepSeek, OpenAI, xAI, Kimi,
Doubao, SiliconFlow, OpenRouter…). Anthropic/Gemini native thinking is
out of scope for this version.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Remove the "Reasoning Model" and "Disable Thinking Mode" toggles from
the model config dialog. These switches were injecting `enable_thinking`
(a Qwen/DashScope-specific request parameter) into all providers,
causing 400 errors on OpenAI, Anthropic, Gemini, and other official
services.
Replace with automatic inference in buildPiModel:
- reasoning: true when the model's catalog capabilities include
'reasoning'; falls back to a model-name heuristic for unlisted
custom/self-hosted models (qwen3, qwq, deepseek-r, o1/o3, etc.).
- compat.thinkingFormat 'qwen': only set when reasoning is true AND
the model is from the qwen provider or the model id contains
"qwen"/"qwq". All other providers get compat: undefined and will
never receive enable_thinking.
This means official reasoning models (o3, deepseek-reasoner, etc.) now
automatically get thinkingLevel:'medium', and self-hosted Qwen3
models continue to work as before — without any manual toggle.
Behavior change: models with 'reasoning' in their catalog capabilities
now default to thinkingLevel:'medium' (previously always 'off' unless
the manual switch was enabled).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Split the monolithic MessageView into two focused modules:
- TypeAnalysisView: type distribution, text depth analysis, length distributions
- TimeAnalysisView: hourly/weekday/monthly/yearly distributions, member trend, heatmap, calendar
Each tab has a redesigned narrative profile card with rich insights:
- TypeProfileCard: date range, total messages, non-text breakdown pills, 4 metric cards
- TimeProfileCard: active days narrative, time personality labels (multi-dimensional), 4 metric cards
Also: reduce responseLatency font size in RelationshipView for better readability.
Co-authored-by: Cursor <cursoragent@cursor.com>
Wraps each argument in double quotes when building the systemd unit's
ExecStart= line, so installs under paths containing spaces (e.g. home
directories with spaces, nvm paths) no longer cause systemd to
mis-tokenize the command and fail to start the service.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Warn users that regeneration will clear all existing AI summaries.
Uses the same UModal pattern as other confirmation dialogs in settings.
Co-authored-by: Cursor <cursoragent@cursor.com>
- insertMessageAfter: update active_message_id when appending to tail
so the new reply is on the active path after reload
- editAndOverwriteAll: add hasConfig + abort checks before
deleteMessagesFrom to prevent data loss when LLM is unconfigured or
a concurrent request starts
- editCurrentRoundOnly: defer all DB mutations until after stream
completes; on abort/error only in-memory state is restored
- sync in-memory parentId after relinking so follow-up edits send the
correct historyLeafMessageId
Co-authored-by: Cursor <cursoragent@cursor.com>
Remove the message branching/tree system in favor of a simpler editing model:
- Default: edit a message and only regenerate the current round's AI response
- Optional: overwrite all subsequent messages when editing
Key changes:
- Backend: add updateMessageContent, deleteAndRelinkMessage, insertMessageAfter,
deleteMessagesFrom, forkConversation; remove createMessageBranch, switchMessageBranch
- Platform layer: add IPC/HTTP handlers for new methods across Electron and CLI
- Frontend store: refactor editMessageAndRegenerate with two modes
- UI: replace branch navigation with "overwrite subsequent" checkbox,
auto-resizing edit textarea
- Fix insertMessageAfter missing debug_context from pendingDebugContextMap
- Update tests to cover new editing methods
Co-authored-by: Cursor <cursoragent@cursor.com>
- Replace 6-subcommand daemon group with simpler top-level design:
chatlab start --daemon install as system service (launchd/systemd)
chatlab stop uninstall service and disable auto-start
chatlab status show resident service status
- Drop temporary background process layer; daemon = system service only
- Store port/host meta in ~/.chatlab/daemon.json for status display
- Fix --no-open description: Web UI still runs, only browser is suppressed
- Update quick-start docs (cn/en/tw) to reflect new commands
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Move quick-start.md to usage/ across cn/en/tw
- Switch locale index pages from home to doc layout
- Add Baidu analytics script and adjust nav doc links
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- Fix apiKeySet reporting false positives for local no-key configs by
resolving the actual key via resolveApiKey and excluding placeholders
- Prevent editing a config from reusing an incompatible API key when
provider or connection mode has changed
- Pass summary strategy to single-session generation for consistency
Co-authored-by: Cursor <cursoragent@cursor.com>
Previously clicking stop would only take effect after the current LLM
request finished. Now the UI state updates instantly while the in-flight
request completes silently in the background.
Co-authored-by: Cursor <cursoragent@cursor.com>
Support two summary strategies (brief/standard) for AI-generated session
summaries. Users can choose their preferred detail level in Settings > AI >
Session Index. Standard (default) produces richer summaries covering people,
events, and key details.
Co-authored-by: Cursor <cursoragent@cursor.com>
- Implement generateSummary/checkCanGenerateSummary in FetchSessionIndexAdapter
via existing HTTP endpoints instead of stubs that returned instantly
- Add checkCanGenerate service function in summary-service.ts
- Add POST /_web/sessions/:id/summaries/check-can-generate route
The page froze because stub methods returned synchronously, causing a tight
loop of hundreds of Vue reactive updates without any network delay.
Co-authored-by: Cursor <cursoragent@cursor.com>
- Add apiKeySet field to GET /_web/ai/llm/configs response
- Allow saving in edit mode without re-entering API key when key already set
- Skip API key validation on save when editing with existing key unchanged
- Show edit-mode placeholder hint on API Key input
- Add missing customModels/disableThinking/isReasoningModel to CLI route types
Co-authored-by: Cursor <cursoragent@cursor.com>
Update links pointing to documentation pages (usage guides, format specs,
troubleshooting, API docs) across README, source code, and i18n files to
use docs.chatlab.fun. Main site links (homepage, roadmap, community,
assets) remain unchanged.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Move SessionIndexSection from StorageTab to AISettingsTab (placed above
export settings), rename label from "会话管理" to "会话索引" across all
locales, swap automation tab before data management tab, and simplify
StorageTab by removing its now-unnecessary SubTabs navigation.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Restructure import components into dedicated import/ directory with new
ImportTabSelector, ApiImportCard, and AutoSyncCard components. Add demo
import, entry animation, and tutorial links to home page. Update i18n
translations for all locales. Add chatApi init guard in registry and
defaultPath to DataDirInfo preload interface.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Replace @mariozechner/pi-ai and pi-agent-core with @earendil-works scope
- Adapt to AgentTool.execute() params type change (Record → unknown)
- Migrate Agent constructor to use initialState for systemPrompt/tools/messages
- Update onPayload callback signature to include model parameter
- Filter finalMessages with isPiMessage() for new AgentMessage union type
- Use state.errorMessage instead of deprecated state.error
Co-authored-by: Cursor <cursoragent@cursor.com>
- Add "type": "module" to all workspace package.json files
- Replace __dirname with import.meta.url-based resolveCliPath() helper
- Convert stream-json/stream-chain CJS named imports to default + destructure
for ESM compatibility
- Use createRequire() for @node-rs/jieba native module loading in ESM context
Co-authored-by: Cursor <cursoragent@cursor.com>
CLI Web was missing chatApi.exportSessionsToTempFiles,
chatApi.cleanupTempExportFiles and the entire mergeApi shim,
causing "not a function" errors when merging sessions in
Settings > Data Management.
Co-authored-by: Cursor <cursoragent@cursor.com>
The unified directory migration (needsUnifiedDirMigration) would
re-trigger after a user changed their data directory, overwriting
config.toml back to the old Electron path.
Root cause: no persistent flag tracking whether the migration had
already completed. If old databases remained (cleanup failure or
timing), it would re-detect and re-run.
Fix:
- Add `electron_migration_done` flag to config.toml [data] section
- Set flag after migrateToUnifiedDirs succeeds and in setCustomDataDir
- Check flag at start of needsUnifiedDirMigration to skip if done
- Add "migrate to default" button when data is at non-default path
- Fix NLP directory path: move from userDataDir to systemDir (~/.chatlab/nlp)
Co-authored-by: Cursor <cursoragent@cursor.com>
CLI Web's Settings > Storage Management was completely non-functional
(stub cacheApi returning empty data). This adds real HTTP endpoints
and wires the frontend to use them.
Backend (cache.ts):
- GET /_web/cache/info — scan databases/ai/cache/logs directory sizes
- POST /_web/cache/clear — clear cache or logs directory (allowlist)
- POST /_web/cache/open-dir — open directory in system file manager
- GET /_web/cache/data-dir — return current data directory path
- GET /_web/cache/latest-import-log — find latest import log file
Frontend:
- Replace cacheApi stubs with real HTTP calls in registry.ts
- Hide data directory migration UI in CLI Web (IS_ELECTRON guard)
- Add "Open Root Dir" button for ~/.chatlab/ in header area
Co-authored-by: Cursor <cursoragent@cursor.com>
Move auto-close logic from finally to success path so the modal
remains open for retry when generate() fails in force mode.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
P1: Small terminal pages with hasMore=false were incorrectly treated as
empty based on file size alone, causing valid messages to be silently
dropped. Now checks actual message content via fileContainsMessages().
P1: Retry import results were not validated, so import failures (including
needFullResync) were masked as success. Now properly handles all failure
cases and tracks retryHasMore to exit the retry loop correctly.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Replace UButton :loading with animate-spin on the arrow-path icon itself.
The sync button's icon now rotates when syncing, and the button is disabled.
The "sync all" button is also disabled when any session is syncing.
50 pages (50k messages) was too low for groups with 100k+ messages.
Raise to 5000 pages (5M messages) as a safety valve while allowing
full-sync to complete in a single pull cycle for most conversations.
After pull-sync imports new messages, generate session index incrementally
instead of full regeneration. This preserves existing sessions and their
summaries, only processing newly imported (unindexed) messages.
Adds generateIncremental method across the full stack:
- core: already had generateIncrementalSessionIndex
- node-runtime: expose generateIncrementalIndex
- CLI: add POST /_web/sessions/:id/generate-incremental-index route
- Electron: add session:generateIncremental IPC handler
- Frontend: add generateIncremental to SessionIndexAdapter interface
After importing messages via pull sync, automatically generate session
index for the affected sessions. This ensures the timeline and session
overview are immediately available without manual index generation.
After a sync finishes (triggerPull/triggerPullAll/onPullResult), reload the
session list so newly imported chats appear in the sidebar immediately.
Also refresh after removing a subscription with deleteData=true so the
deleted session disappears from the sidebar.
When the remote server doesn't return a `page` field and the result count
equals the requested limit, automatically re-fetch with limit=5000 to get
the complete session list instead of showing truncated results.
- Show confirmation dialog with two choices: keep data or delete imported chats
- Backend supports deleteData param across CLI routes and Electron IPC
- DataSourceManager.removeSession now returns the removed session info
- Add i18n strings for the delete confirmation dialog (zh-CN/en-US/ja-JP/zh-TW)
- Add multi-retry (3 attempts, 2s/3s/5s delays) for cold-start empty responses
- Remove unused offset/end parameters from fetch URL construction
- Fix Accept header to application/json only (avoids content negotiation issues)
- Add PullProgress type and progress tracking in PullEngine
- Export progress via getProgress() for frontend polling
sessionExists() previously only checked file existence, causing pull sync
to enter incremental import path on invalid/empty DB files. Now validates
that the message table exists, and maps schema errors to needFullResync
as a safety fallback.
Add a hover-triggered CaptureButton to each expanded changelog entry so
users can screenshot individual version notes. Make the @zumer/snapdom
markdown list compatibility fix opt-in via a new `markdownFix` option
(default false), so custom-styled lists like changelog bullets are no
longer disrupted during capture.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Extract version/tag/prerelease detection into a shared prepare_release job,
split CLI packaging into its own build-cli artifact, switch publish-cli to
download the pre-built artifact instead of rebuilding inline, pin pnpm install
with --frozen-lockfile, and fix changelog summary lookup to match by version
instead of always selecting the first entry.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
When a pullAllSessions was already running, the reloadTimer triggered
by adding sessions would be skipped by the source-level concurrency
lock, causing newly added sessions to wait until the next scheduled
interval (up to 60 minutes) before syncing.
Fix: trigger individual per-session pulls via triggerPull(sourceId,
sessionId) which bypasses the source-level lock, ensuring immediate
sync for newly added sessions regardless of concurrent pulls.
Also includes: connection pre-check on data source add, removal of
duplicate "browse remote sessions" button, hiding "other" session
types, and full resync fix for deleted local sessions.
Co-authored-by: Cursor <cursoragent@cursor.com>
Extract shared business logic from monolithic web.ts into reusable
services in @openchatlab/node-runtime (session, member, index, summary,
export), using SessionRuntimeAdapter to abstract platform differences.
Split web.ts into 12 domain-specific route modules under routes/web/.
Co-authored-by: Cursor <cursoragent@cursor.com>
MiniMax models embed thinking content as <think> tags in the content
field rather than using a separate reasoning field. Add a streaming
think-tag parser that detects these tags in real-time during streaming
and routes them to the thinking UI instead of displaying raw tags.
Co-authored-by: Cursor <cursoragent@cursor.com>
Add directory import support for multi-file formats (e.g. shuakami chunked-jsonl).
Users can select or drag-and-drop a directory via the "Import Folder" checkbox.
- Add `findEntryFileInDirectory()` to parser layer for format-agnostic entry file discovery
- Add `importDirectory` method to ImportAdapter interface
- Implement FetchImportAdapter.importDirectory with multi-file FormData upload
- Add `POST /_web/import-directory` CLI route with directory reconstruction
- Add `chat:importDirectory` IPC handler in Electron main process
- Enhance FileDropZone with `directory` prop and drag-and-drop directory support
- Refactor FetchImportAdapter to extract reusable `consumeSseStream` helper
Co-authored-by: Cursor <cursoragent@cursor.com>
The IncrementalImportModal was Electron-only: showOpenDialog returned
canceled in web mode and FetchImportAdapter threw notSupported for both
analyzeIncrementalImport and incrementalImport despite CLI routes
already existing. Fix by using FileDropZone's built-in file input for
file selection and implementing the two fetch adapter methods via
FormData upload + SSE streaming.
Co-authored-by: Cursor <cursoragent@cursor.com>
diguaandCursor
