home / mcp / telegram mcp server

Telegram MCP Server

Telegram MCP for managing dialogs, messages, drafts, read statuses, and more.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "chaindead-telegram-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "@chaindead/telegram-mcp"
      ],
      "env": {
        "TG_APP_ID": "<your-api-id>",
        "TG_API_HASH": "<your-api-hash>"
      }
    }
  }
}

You connect your AI assistant to Telegram through a lightweight MCP server that acts as a bridge between the Telegram API and local tools. This lets you manage dialogs, read messages, and draft responses from your AI client while keeping control over how data flows between Telegram and your applications.

How to use

You interact with the Telegram MCP server from your MCP client (for example Claude Desktop or Cursor). Start the server using one of the supported installation methods, then configure your client to recognize the Telegram MCP server so your AI assistant can manage dialogs, read messages, and send drafts.

How to install

Prerequisites: a working Node.js environment with npm or npx available, or access to a prebuilt binary if you choose a release installation. You will run a command to connect to the Telegram API and then place the server under your MCP client’s configuration.

# Install via NPX (runs directly without a local installation)
npx -y @chaindead/telegram-mcp

From Releases

You can download prebuilt binaries for your platform and run them directly. Use the appropriate archive for your OS and architecture, then extract and place the binary in a directory in your PATH.

From Source

If you prefer building from source, install Go (for running the server) and then build the MCP server binary from the source code. The project provides a Go-based workflow for building and running locally.

Available tools

tg_me

Retrieve current Telegram account information such as username and phone number through the Telegram MCP endpoint.

tg_dialogs

List dialogs (conversations) with optional unread filter to understand your active chats.

tg_read

Mark dialogs as read to keep your message state synchronized.

tg_dialog

Retrieve messages from a specific dialog, enabling analysis and drafting based on context.

tg_send

Send draft messages to chosen dialogs, allowing you to compose and dispatch replies via the MCP bridge.