wa-chat-reader
Reads your local WhatsApp chat history and exposes tools to retrieve messages and chat lists, enabling querying and summarization of your conversations entirely offline.
README
wa-chat-reader
<!-- mcp-name: io.github.NabheetCloud/wa-chat-reader -->
Unofficial, read-only MCP server for your local WhatsApp chat history.
wa-chat-reader reads the WhatsApp ChatStorage.sqlite database that the WhatsApp
desktop app keeps on your machine and exposes two tools to any MCP client (Claude
Code, Claude Desktop, etc.) so you can summarize, search, and reason over your own
messages. Everything runs locally — no data ever leaves your computer.
⚠️ Not affiliated with, endorsed by, or connected to WhatsApp or Meta. "WhatsApp" is a trademark of Meta Platforms, Inc. This is an independent, unofficial tool that only reads a database WhatsApp already stores on your device.
Privacy & security
This tool touches highly sensitive data — your private conversations. Read this first.
- Read-only. The database is opened with SQLite
mode=ro&immutable=1; the server never writes to, modifies, or deletes anything. - Local-only. There is no network code. Messages are returned to your MCP client and nowhere else. Whatever LLM/client you connect will, of course, see the message text you ask it to read — you control that by choosing which client to run.
- Your consent, your data. Only run this against your own device and your own chats. Do not use it to access anyone else's messages.
- macOS access. The WhatsApp database lives in a protected group container. The
process running this server needs Full Disk Access
(System Settings → Privacy & Security → Full Disk Access), or point
WHATSAPP_DB_PATHSat a readable copy you export yourself. - Scope control.
WHATSAPP_MAX_DAYS(default 30) caps how far back any single call can read.
Install
Requires Python ≥ 3.10. Recommended: uv so the server
runs in an isolated environment.
# one-shot run (no install)
uvx wa-chat-reader
# or install with pip
pip install wa-chat-reader
wa-chat-reader
Use in Claude Code
claude mcp add wa-chat-reader -- uvx wa-chat-reader
Or install as a plugin (see the marketplace section below), then ask Claude things like:
- "Summarize my unread WhatsApp group chats from the last 3 days."
- "List the WhatsApp chats I've been active in this week."
Configuration
All configuration is via environment variables (all optional):
| Variable | Default | Description |
|---|---|---|
WHATSAPP_DB_PATHS |
platform default | ;-separated candidate paths to ChatStorage.sqlite (first readable wins) |
WHATSAPP_BACKUP_PATHS |
— | ;-separated fallback paths, tried after the primary paths |
WHATSAPP_MAX_DAYS |
30 |
Max look-back window (days) a caller may request |
WHATSAPP_DEBUG |
false |
true for verbose logging to stderr |
WHATSAPP_LOG_FILE |
— | Optional path to also write logs to a file |
Default database locations:
| OS | Path |
|---|---|
| macOS | ~/Library/Group Containers/group.net.whatsapp.WhatsApp.shared/ChatStorage.sqlite |
| Windows | ~/AppData/Roaming/WhatsApp/Databases/ChatStorage.sqlite |
| Linux | ~/.local/share/whatsapp/ChatStorage.sqlite |
Tools
| Tool | Description |
|---|---|
get_whatsapp_messages |
Concatenated message text from group and personal chats within a look-back window. Params: days (1–max), chat_type (all/groups/personal). Read-only. |
get_whatsapp_chat_list |
Active chat names (no message bodies) within a look-back window. Param: days. Read-only. |
Both tools are annotated readOnlyHint: true.
Bonus: activity dashboard
whatsapp_dashboard.py renders a self-contained, offline HTML dashboard of your
messaging activity (volume over time, daily/weekly rhythm, top chats) straight from
the same database:
python3 whatsapp_dashboard.py # -> whatsapp_dashboard.html
The generated HTML contains real contact names — it is git-ignored and meant to stay local.
Develop
git clone https://github.com/NabheetCloud/wa-chat-reader
cd wa-chat-reader
pip install -e .
python3 simple_whatsapp_mcp.py # run from source without installing
License
MIT © 2026 Nabheet Madan
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。