tg-mcp / site
llms.txt2 starsChanged 10 days agoMITView on GitHub
tg-mcp is a free, open-source (MIT), unofficial remote MCP server for a personal Telegram account. Add https://connector.tgmcp.dev/mcp to Claude, ChatGPT, Gemini, Claude Code, Cursor or any MCP client that speaks streamable HTTP, sign in by scanning a QR code with Telegram, and the AI can read and search your chats. It writes only if you allow it at sign-in: Drafts you send yourself, or messages to chats you pick, sent at once within the server's limits (your AI app decides…
In this repository28 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- ask-mattSkill
- codebase-designSkill
- code-reviewSkill
- diagnosing-bugsSkill
- domain-modelingSkill
- grillingSkill
- grill-meSkill
- grill-with-docsSkill
- handoffSkill
- implementSkill
- implement-specSkill
- improve-codebase-architectureSkill
- prototypeSkill
- prSkill
- researchSkill
- retroSkill
- setup-matt-pocock-skillsSkill
- tddSkill
- teachSkill
- to-questionnaireSkill
- to-specSkill
- to-ticketsSkill
- triageSkill
- wait-whatSkill
- wayfinderSkill
- wizardSkill
- writing-for-agentsSkill
The file
We split this llms.txt into the rules it gives an agent
3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 5
Not affiliated with Telegram.
- Distinctive: line 21
Read: list chats (unread only, by Chat folder, archived), read messages and threads, search messages across chats, find chats and contacts …
- Distinctive: line 24
Never: edit, delete or forward messages
≈ “Common” is an estimate: at least 1 in 5 of the 3,727 llms.txt files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
# tg-mcp > tg-mcp is a free, open-source (MIT), unofficial remote MCP server for a personal Telegram account. Add `https://connector.tgmcp.dev/mcp` to Claude, ChatGPT, Gemini, Claude Code, Cursor or any MCP client that speaks streamable HTTP, sign in by scanning a QR code with Telegram, and the AI can read and search your chats. It writes only if you allow it at sign-in: Drafts you send yourself, or messages to chats you pick, sent at once within the server's limits (your AI app decides whether to ask you first). Not affiliated with Telegram. Made by Kirill Bobykin (https://github.com/qelphybox), an independent developer. In Russian: https://tgmcp.dev/ru/ (the same page, the same server). ## Connect - Public server URL: `https://connector.tgmcp.dev/mcp` (streamable HTTP, OAuth discovered automatically, no API key, no sign-up, no payment) - Claude (web, desktop, mobile; free plan too): Settings → Connectors → Add custom connector → paste the URL → Connect → scan the QR code with Telegram on your phone (Settings → Devices → Link Desktop Device) - ChatGPT (paid plans): Settings → Apps & Connectors → Advanced → turn on Developer mode → Create → name it Telegram, paste the URL → scan the QR code - Gemini (personal Google accounts, add it on the web): gemini.google.com → Settings → Connected apps → Add a custom app → paste the URL → approve → scan the QR code - Claude Code: `claude mcp add --transport http telegram https://connector.tgmcp.dev/mcp`, then `/mcp` → telegram → Authenticate - Cursor and other MCP clients: add the URL as a streamable-HTTP server - Disconnect: Telegram → Settings → Devices → end the "Unofficial MCP Server" session ## What the AI can do - Read: list chats (unread only, by Chat folder, archived), read messages and threads, search messages across chats, find chats and contacts by name, read chat and profile info, list group members, read Drafts - Media: look at photos and video thumbnails, transcribe voice messages and video notes with Telegram's own transcription, or download files through a short-lived link - Write, only when allowed at sign-in: `save_draft` (text in a chat's message field that you send yourself); `send_message`, `react` and `mark_read` in the chats you picked, sent at once (your AI app decides whether to ask you first; the server enforces the chats and the limits) - Never: edit, delete or forward messages ## Docs - [README](https://github.com/qelphybox/tg-mcp/blob/main/README.md): quick start, full tool reference, prompts, security model, self-hosting - [Tools](https://github.com/qelphybox/tg-mcp#tools): every tool with its inputs and outputs - [Security model](https://github.com/qelphybox/tg-mcp#security-model): the Telegram session is stored encrypted with a key that only the AI app's tokens carry; message content is never stored or logged - [Self-hosting](https://github.com/qelphybox/tg-mcp#self-hosting): one Docker image for amd64 and arm64 (`ghcr.io/qelphybox/tg-mcp`) that runs with no configuration and a built-in Telegram app; `compose.yaml` adds HTTPS through Caddy for claude.ai and Claude desktop - [Privacy policy](https://tgmcp.dev/privacy.html): what the public server stores and what it never stores ## Optional - [Architecture decisions](https://github.com/qelphybox/tg-mcp/tree/main/docs/adr) - [Releases](https://github.com/qelphybox/tg-mcp/releases) - [Issues and support](https://github.com/qelphybox/tg-mcp/issues)
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

