tg-mcp-tools
A local MCP server for Cursor and Claude Desktop that reads Telegram subscriptions over MTProto and exposes channel posts for digests and natural-language queries.
README
tg-mcp-tools
A local MCP server for Cursor and Claude Desktop. It reads your Telegram subscriptions over MTProto (personal account, not a bot) and exposes channel posts to the agent for digests and natural-language questions like “what’s new in my Travel folder this week?”
Install
npm install -g tg-mcp-tools
Or run without installing (recommended for MCP):
npx tg-mcp-tools-auth # one-time login
npx tg-mcp-tools # MCP server (stdio)
Features
- 13 MCP tools: channels (create/rename), folders (create/edit), dialogs, messages, digests, send message, archive/unarchive
- Post links — every message includes a
urlfield (https://t.me/...) - Archived channels — included when listing subscriptions
- QR login — scan from the terminal, no SMS or in-app codes
Requirements
- Node.js 18+
- Telegram API credentials from my.telegram.org/apps
Quick start
Create a project directory, add credentials, authorize, then connect MCP.
mkdir my-telegram-mcp && cd my-telegram-mcp
cp path/to/node_modules/tg-mcp-tools/.env.example .env
# or from a git clone: cp .env.example .env
Fill in .env:
TELEGRAM_API_ID=12345678
TELEGRAM_API_HASH=your_api_hash
Authorize (scan QR in terminal; session is saved to .env in the current directory):
npx tg-mcp-tools-auth
# or, from a git clone: npm run auth
Verify the MCP server starts:
npx tg-mcp-tools
# or, from a git clone: npm run mcp
Cursor setup
Add to .cursor/mcp.json in your project (where .env lives):
{
"mcpServers": {
"telegram": {
"command": "npx",
"args": ["-y", "tg-mcp-tools"]
}
}
}
The server loads .env from the current working directory (your project root). You don't need to duplicate env vars in mcp.json if .env is present.
For local development from a git clone, use:
{
"mcpServers": {
"telegram": {
"command": "npm",
"args": ["run", "mcp"]
}
}
}
After changing the config, reload Cursor (Developer: Reload Window).
MCP Inspector
Use the MCP Inspector to call tg_* tools in the browser and inspect JSON-RPC traffic before wiring Cursor.
From the project root (where .env lives):
npm run inspect
This builds the server, starts it via npm run mcp over stdio, and opens the Inspector UI (usually http://localhost:6274). Use the full URL printed in the terminal if an auth token is required.
Links
CLI mode (no browser):
npx -y @modelcontextprotocol/inspector --cli npm run mcp
MCP tools
| Tool | Purpose |
|---|---|
tg_create_channel |
Create a new broadcast channel (title and optional description) |
tg_set_channel_title |
Rename a channel or group (requires admin rights) |
tg_get_folders |
Telegram chat folders (tabs) with id and name |
tg_create_folder |
Create a folder with channels and filter rules |
tg_edit_folder |
Update an existing folder: replace included chats or rename it |
tg_get_dialogs |
List channels, groups, and chats; optional folder filter |
tg_get_messages |
Posts from one channel (@username or numeric id) |
tg_get_recent_from_channels |
Digest from multiple channels, sorted by date |
tg_get_recent_from_folder |
Digest from all channels/groups in a folder |
tg_send_message |
Send formatted text to a user (@username), chat id, or "me" (Markdown/HTML) |
tg_archive_chats |
Move chats/channels to Archive |
tg_unarchive_chats |
Restore chats/channels from Archive |
Example prompts in Cursor:
What’s new in my Travel folder this week?
Authorization
- Run
npx tg-mcp-tools-authfrom the directory where.envshould live - On your phone: Telegram → Settings → Devices → Link Desktop Device
- Scan the QR code in the terminal
- Enter your 2FA password if prompted
TELEGRAM_SESSIONis written to.envautomatically
Scripts
| Command | Description |
|---|---|
npx tg-mcp-tools-auth |
One-time login; saves TELEGRAM_SESSION to .env in cwd |
npx tg-mcp-tools |
Start the MCP server (stdio) |
npm run auth / npm run mcp |
Start auth or MCP server (npm run build first when developing from a git clone) |
npm run inspect |
Build and launch MCP Inspector for this server |
npm run build |
Compile TypeScript → dist/ |
Security
- Do not commit
.env— it contains your Telegram session (full account access) - Never share
TELEGRAM_SESSIONin logs, issues, or chats - Revoke API credentials or sessions at my.telegram.org
Project layout
src/
├── config.ts # load .env from repo root
├── env-file.ts # update .env values after auth
├── telegram-client.ts # mtcute: dialogs, history, post URLs
├── index.ts # MCP server and tool registration
└── auth.ts # CLI auth (QR)
Stack
- @mtcute/node — MTProto client
- @modelcontextprotocol/sdk — MCP
- TypeScript (ESM), zod, dotenv
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.