Telegram MCP: No Official Server, and the Ban Risk
No — Telegram does not have an official MCP server. Telegram hasn’t published anything for the Model Context Protocol. What you’ll find searching “Telegram MCP” are community-built servers, and they split into two very different camps: ones that use Telegram’s sanctioned Bot API (a bot token from @BotFather) and ones that log into your personal account through MTProto — the second kind acts as you and can get an account flagged or banned. Neither is something Telegram built or endorses.
Even setting that aside, the same limit applies to any of them: an MCP server hands Telegram to an AI inside a conversation you start. It’s a doorway, not a worker. Nothing watches your chats for you, nothing fires when a message arrives or a keyword shows up, and nothing runs while the chat is closed. Here’s the honest state of Telegram and MCP, what the community servers actually do, where the limits are — and what to use when you want Telegram work that runs on its own.
Looking beyond Telegram? MCP servers by category covers the rest.
What an MCP connection to Telegram does
Model Context Protocol (MCP) is the open standard that lets an AI client — Claude, ChatGPT, Cursor, and others — talk to an outside app through a shared interface. Telegram hasn’t built one. The servers people mean when they say “Telegram MCP” are independent projects, and which one you pick decides what it can reach:
- Bot API servers — like
guangxiangdebizi/telegram-mcp— use a bot token from @BotFather. A bot can send messages, share files and media, forward, and delete within chats it’s been added to. It’s the sanctioned path, but a bot can’t read your personal DMs or your full chat history. - MTProto / userbot servers — like
chigwell/telegram-mcp,dryeab/mcp-telegram(both built on Telethon),sparfenyuk/mcp-telegram(read-only MTProto), andmcp-telegram/mcp-telegram(GramJS) — log in as your personal account. They get your real chats, contacts, and message history, and can send, edit, delete, forward, pin, and search as you. Some, likeantongsm/mcp-telegram, support both modes.
With one connected, an AI client can typically:
- Read chats and messages — pull a conversation, a channel’s recent posts, or search across your chats into a chat window.
- Send and manage messages — send a message, reply, forward, edit, or delete, using either the bot’s identity or (with a userbot) your own account.
- Look up contacts and groups — list chats, find a contact, or get info about a group or channel.
It’s genuinely useful for ad-hoc work — “summarize the last 50 messages in my project group” — answered from live Telegram data instead of a guess.
How to connect Telegram to AI in 2026
There’s no “install the official server” step, because Telegram didn’t build one. Which path you take depends on the camp above:
- For the Bot API path, message @BotFather in Telegram, create a bot, and copy its token. Add the bot to whatever chats or channels you want it to act in — it only sees those. This is the low-risk route.
- For the MTProto path, register a Telegram application at my.telegram.org to get an
api_idandapi_hash, then log the server in with your phone number. This gives full access to your account — and full account risk with it. - Add the community MCP server to your AI client, point it at your token or API credentials, and restart the client so the tools register.
- Test with a read query first — pull one chat’s recent messages — before trusting any server with send, delete, or forward tools, especially a userbot acting as your own account.
For most people searching “connect Telegram to AI,” this is a real setup with a real decision behind it, not a one-click connector.
Where the Telegram MCP stops
Even setting the “bot or userbot, and which repo do I trust” question aside, the same four limits show up here:
- It only works inside a chat you start. Close the window and nothing happens. Nothing is watching your Telegram chats or a channel for you.
- No triggers. A new message arriving, a keyword mentioned in a group, a file dropped in a channel — none of these can start anything through MCP. There’s no “when this happens in Telegram, do that.”
- It’s one app at a time. A Telegram MCP server knows Telegram. Getting a message into your CRM, a task list, and a spreadsheet means wiring up a separate MCP server for each and hoping your client can juggle them in one turn.
- You own the plumbing — and the account risk. There’s no vendor support line here. With a userbot, you registered the app, and Telegram throttles and may flag or ban accounts that automate through the user API. A server with send tools acts as your account with no one else reviewing what it does.
So a Telegram MCP server, community-built as all of them are, is not a way to make Telegram run — to have work happen on a schedule or in reaction to an event, across the other tools a message touches.
Running Telegram work that doesn’t need a chat open
That “run on its own, across apps” gap is exactly where Carly fits. Carly connects to Telegram and to the ~260 other apps it supports natively, plus anything with a public API through your own key — no community MCP server to vet, no personal-account userbot to risk. The difference from MCP is the important part: Carly’s workflows are triggered and scheduled, so Telegram work happens whether or not anyone has a chat window open.
A few things a Carly workflow can do that no MCP server can:
- When a message hits a Telegram channel you monitor → extract the details, log them to a Google Sheet, and post a heads-up in Slack — automatically, the moment it arrives.
- Every morning → send a Telegram message with your day’s calendar and the tasks due, so it’s waiting when you wake up.
- When a lead fills out a form → create the contact in your CRM and send your team a Telegram alert with the details.
The non-AI steps — the moving, matching, and routing between apps — are free and unlimited, the Zapier-style backbone of the workflow. The AI steps (drafting, summarizing, deciding) start at $35/month. You describe the outcome in plain language and Carly wires up the Telegram connection and everything downstream.
If you just want to ask questions about a chat from a conversation, a community MCP server is the closest thing available, with a bot token or a personal-account login behind it. If you want Telegram activity to actually trigger something — on a schedule, in reaction to an event, across every app that message touches — that’s the job no Telegram MCP server was built for.
FAQ
Does Telegram have an official MCP server? No. Telegram hasn’t published one. What exists are community projects — some built on the sanctioned Bot API (a token from @BotFather), others logging into your personal account through MTProto with Telethon or GramJS.
Is it safe to connect Telegram to AI over MCP? It depends which kind. A Bot API server is low-risk — the bot only acts in chats you add it to. A userbot server that logs into your personal account is riskier: Telegram throttles and can flag or ban accounts that automate through the user API, and the server acts as you.
Can a Telegram MCP server trigger automations? No. MCP is request/response inside an AI chat — it has no triggers and nothing runs when the conversation is closed. For event- or schedule-driven Telegram work across apps, you need a workflow tool like Carly rather than an MCP server.
Can I connect Telegram to AI without coding or hosting a server? Yes. You don’t have to register a Telegram app, manage a bot token, or run an MCP server at all. Carly connects to Telegram for you and lets you build the automation in plain language — describe what you want to happen and it wires up the Telegram connection and the other apps involved, with no server to host and no code to write.
Ready to automate your busywork?
Carly schedules, researches, and briefs you—so you can focus on what matters.
See what people say
"Before Carly, I relied on a Calendly link, but the whole process felt impersonal and not very professional. Carly changed that by handling all the back-and-forth, so I'm no longer stuck in endless email threads trying to line up schedules.
Now Carly reaches out to candidates, shares my real-time availability, lets them pick a slot, then sends a Zoom link and drops it straight into my calendar. She sends reminders to both of us before each call, which has significantly reduced no-shows and last-minute confusion.
On top of scheduling, Carly acts like a full executive assistant, sending me my schedule the night before so I can prepare for each call. It reminds me of the old x.ai assistant, but Carly is noticeably smarter, faster, and better suited to my healthcare recruitment business."
