Telegram-Client

Telegram-Client

Bridge between the Telegram API and AI assistants via the Model Context Protocol (MCP).

243
Stars
28
Forks
20
Releases

Overview

Telegram MCP server is a bridge between the Telegram API and AI assistants built on the Model Context Protocol. It enables MCP-enabled agents to access and manage Telegram data by exposing Telegram operations as MCP tools: get current account information (tg_me), list dialogs with optional unread filter (tg_dialogs), mark dialog as read (tg_read), retrieve messages from a specific dialog (tg_dialog), and send draft messages to any dialog (tg_send). The server is OS-agnostic and can be installed via Homebrew, NPX, releases, or built from source. It requires Telegram API authentication (API ID, API HASH, and phone number) and supports 2FA via an optional password. Prompt examples illustrate Message Management, Organization, and Communication workflows. The MCP server emphasizes compliance with Telegram’s Terms of Service and provides integration guidance for Claude Desktop and Cursor through client configuration. This setup enables AI assistants to query Telegram data and draft responses through MCP, facilitating safe and controlled interactions with Telegram data via MCP.

Details

Owner
chaindead
Language
Go
License
MIT License
Updated
2025-12-07

Features

Get current account information (tool: tg_me)

Expose the Telegram account context to MCP-enabled AI assistants by retrieving current account information.

List dialogs with optional unread filter (tool: tg_dialogs)

Provide a paged list of user dialogs, with optional filtering for unread messages to aid prioritization.

Mark dialog as read (tool: tg_read)

Update read status for specified dialogs to reflect user activity and keep context in sync.

Retrieve messages from specific dialog (tool: tg_dialog)

Fetch messages from a particular Telegram dialog for analysis, processing, or drafting within MCP.

Send draft messages to any dialog (tool: tg_send)

Create and send draft messages to dialogs on behalf of the user, enabling composed responses via MCP.

Tags

TelegramMCPModel Context ProtocolDialog managementMessage retrievalRead statusDraft messagesAI assistants