diguaandClaude Fable 5
9c7778f7e3
Message-retrieval tools mirror rawMessages inside data, and the preprocessing pipeline rendered that array into the tool result text as "rawMessages: [object Object], [object Object], ..." — roughly 1.7KB of junk per call (~400-500 wasted tokens) that also polluted the truncated result text persisted for history replay. - formatToolResultAsText skips the rawMessages key (raw message objects are already rendered via the formatted messages list) - CLI tool adapter strips rawMessages from extraDetails before invoking the pipeline, matching the existing desktop adapter behavior Regression tests fail without the fix at both the format layer and the CLI adapter layer. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Your chat history, finally yours.
English | 简体中文
Official Website · Docs · Quick Start · Roadmap · Releases
ChatLab is an open-source desktop app for understanding your social conversations. It combines a flexible SQL engine with AI agents so you can explore patterns, ask better questions, and extract insights from chat data, all on your own machine.
Currently supported: WhatsApp, LINE, WeChat, QQ, Discord, Instagram, and Telegram. Coming next: iMessage, Messenger, and KakaoTalk.
New install? Start here: Getting started
Core Features
- 🚀 Built for large histories: Stream parsing and multi-worker processing keep imports and analysis responsive, even at million-message scale.
- 🔒 Private by default: Your chat data and settings stay local. No mandatory cloud upload of raw conversations.
- 🤖 AI that can actually operate on data: Agent + Function Calling workflows (24+ tools) can search, summarize, and analyze chat records with context.
- 📊 Insight-rich visual views: See trends, time patterns, interaction frequency, rankings, and more in one place.
- 🧩 Cross-platform normalization: Different export formats are mapped into a unified model so you can analyze them consistently.
Installation
Desktop App
Download the installer for your OS from the official website or GitHub Releases, then double-click to install.
CLI
Requires Node.js ≥ 20.
npm i chatlab-cli -g
Start ChatLab:
chatlab start # Start API + Web UI, auto-open in browser
chatlab start --no-open # Start API + Web UI, skip auto-open
chatlab start --headless # API only, no Web UI (for scripts / AI Agents)
Common options: --port <port> (default 3110), --host <address>, --token <token>.
To run as a persistent background service (auto-start on login + auto-restart on crash):
chatlab start --daemon # Install as system service (macOS / Linux)
chatlab status # Check service status
chatlab stop # Stop and uninstall service
For a full walkthrough, see the Quick Start guide.
Usage Guides
Preview
For more previews, please visit the official website: chatlab.fun
Architecture Overview
ChatLab is a pnpm monorepo built on Electron + Vue 3 + Nuxt UI + Tailwind CSS. Core business logic lives in shared packages (@openchatlab/core, @openchatlab/node-runtime, @openchatlab/tools), consumed by both the desktop app and the CLI service — so they stay in sync.
Data flows in five stages: format detection → stream parsing → local persistence → SQL + AI query → visualization.
For a deep dive, see the architecture documentation.
Architecture Principles
- Local-first by default: Raw chat data, indexes, and settings remain on-device unless you explicitly choose otherwise.
- Streaming over buffering: Stream-first parsing and incremental processing keep large imports stable and memory-efficient.
- Composable intelligence: AI features are assembled through Agent + Tool Calling, not hard-coded into one model path.
- Schema-first evolution: Import, query, analysis, and visualization share a consistent data model that scales with new features.
Local Development
For complete contributor instructions, see the Development Guide.
Requirements
- Node.js >= 24 < 25
- pnpm >= 9 < 10
Setup
# Install dependencies
pnpm install
# Run desktop app in dev mode
pnpm dev
# Run CLI in dev mode (apps/cli)
cd apps/cli && pnpm dev
If Electron encounters exceptions during startup, you can try using electron-fix:
npm install electron-fix -g
electron-fix start
Privacy Policy & User Agreement
Before using this software, please read the Privacy Policy & User Agreement.
Community
Please follow these principles before submitting a Pull Request:
- Obvious bug fixes can be submitted directly.
- For new features, please submit an Issue for discussion first; PRs submitted without prior discussion will be closed.
- Keep one PR focused on one task; if changes are extensive, consider splitting them into multiple independent PRs.
- For local setup, repository structure, checks, and AI collaboration notes, see the Development Guide.
Thanks to all contributors:
License
AGPL-3.0 License