Telegram MCP Server
Enables Claude Code to send messages to and receive instructions from Telegram, with task tracking and persistent storage.
README
Telegram MCP Server for Claude Code
A bidirectional Telegram bridge that allows Claude Code to communicate with you via Telegram. Built on the Model Context Protocol (MCP).
Features
- Send messages from Claude Code to your Telegram
- Receive instructions — send Telegram messages that Claude Code can read
- Task tracking — create, update, and monitor tasks with Telegram notifications
- Persistent storage — all messages and tasks stored in SQLite
- Webhook-based — no polling, instant message delivery
Architecture
You (Telegram) ←→ Telegram Bot API ←→ Webhook Server (Express, port 3100) ←→ SQLite DB
↕
Claude Code ←→ MCP Server (stdio) ←→ SQLite DB
Two processes share the same SQLite database:
- Webhook Server — receives your Telegram messages via bot webhook
- MCP Server — spawned by Claude Code via stdio, reads/writes the same DB
Prerequisites
- Node.js 18+
- A Telegram bot (create one via @BotFather)
- Your Telegram chat ID (use @userinfobot to find it)
- A publicly accessible URL for the webhook (e.g., via reverse proxy or tunnel)
Installation
git clone https://github.com/xmart2k/telegram-mcp.git
cd telegram-mcp
npm install
npm run build
Configuration
Copy the example environment file and fill in your values:
cp .env.example .env
Edit .env:
TELEGRAM_BOT_TOKEN=your_bot_token_from_botfather
TELEGRAM_CHAT_ID=your_telegram_chat_id
WEBHOOK_PORT=3100
WEBHOOK_HOST=0.0.0.0
DB_PATH=/path/to/data/telegram-mcp.db
Claude Code Setup
Add the MCP server to your Claude Code configuration (.mcp.json in your project or ~/.claude/settings.json globally):
{
"mcpServers": {
"telegram-pm": {
"command": "node",
"args": ["/path/to/telegram-mcp/dist/mcp-server.js"],
"env": {
"TELEGRAM_BOT_TOKEN": "your_bot_token",
"TELEGRAM_CHAT_ID": "your_chat_id",
"DB_PATH": "/path/to/telegram-mcp/data/telegram-mcp.db"
}
}
}
}
Running the Webhook Server
Start the webhook server to receive Telegram messages:
# Development
npm run dev:webhook
# Production
npm run start:webhook
Systemd Service (optional)
For production, install as a systemd service:
# Edit telegram-webhook.service to match your paths
sudo cp telegram-webhook.service /etc/systemd/system/
sudo systemctl daemon-reload
sudo systemctl enable --now telegram-webhook
Register the Webhook
After the server is running and accessible via a public URL, register the webhook with Telegram (one-time setup):
curl -X POST http://localhost:3100/webhook/setup \
-H "Content-Type: application/json" \
-d '{"url": "https://your-domain.com/webhook"}'
Verify webhook status:
curl http://localhost:3100/webhook/info
MCP Tools
| Tool | Description |
|---|---|
send_message |
Send a Telegram message to you |
get_instructions |
Check for pending messages from you |
acknowledge_instruction |
Mark a specific instruction as processed |
acknowledge_all_instructions |
Mark all pending instructions as processed |
create_task |
Create a tracked task with optional notification |
update_task_status |
Update task status with optional notification |
get_task_status |
Get status and history of a task |
list_active_tasks |
List all non-completed tasks |
list_all_tasks |
List all tasks including completed |
Usage Examples
Once configured, Claude Code can use the tools naturally:
- "Send me a message on Telegram saying the deploy is done" — uses
send_message - "Check if I sent any instructions" — uses
get_instructions - "Create a task BUG-42 to fix the login issue" — uses
create_task - "Mark BUG-42 as completed" — uses
update_task_status
You can also send messages from Telegram to Claude Code — they'll be queued and available via get_instructions.
API Endpoints (Webhook Server)
| Method | Path | Description |
|---|---|---|
POST |
/webhook |
Telegram bot updates (configured automatically) |
POST |
/webhook/setup |
Register webhook URL with Telegram |
GET |
/webhook/info |
Current webhook status |
GET |
/health |
Health check |
License
MIT
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。