CommsConnector
Telegram
Read your Telegram conversations and reply as yourself, in Claude or ChatGPT, through Tulina's own MCP server — with a hosted login, no bot account required.
Official site: telegram.org
What it can do
2 read/1 send
- 01Read
List your conversations and read a thread
- 02Send
Send a message, or reply in an existing thread
- 03Read
Summarize a thread
Connected through a hosted login, a phone number and then the code, with no cookie to paste and no extension to install. One structural difference matters if you are writing a workflow that walks conversations: Telegram has no folders or inboxes. The thread list is flat, where LinkedIn's is not, so "go through my unread" means something different here.
Does Telegram have an MCP server?
We found none from Telegram as of October 2026. The servers that rank for this search are community projects. Some log in as you, through the Telegram client library, to read chats and send messages. Others use a bot account through the Bot API, so they only see chats a bot has been added to. Tulina's connector reads and replies as yourself, through a hosted login run by Unipile, our messaging provider, and needs no bot.
How do you connect Telegram to Claude and ChatGPT?
Through Tulina's MCP server, in three steps:
- Make sure a Unipile subscription key is on the Unipile account connector, either your own key or the platform's, which an admin grants as the hosted messaging option.
- On Tulina's Telegram connector, choose "Connect my Telegram account" and log in with your number and the code Telegram sends you.
- Add
https://mcp.tulina.ai/mcpto Claude or ChatGPT and sign in with your Tulina account. The MCP page has the same link.
Can Claude be integrated with Telegram?
Yes, through any of the routes above. With Tulina, Claude or ChatGPT calls two tools. The first lists your conversations newest first, one to 100 at a time, group chats included and named, and reads the messages of a thread. The second sends. They are separate so that allowing reads never allows sending. A reply goes into an existing thread by its id, or opens a new one with a contact. It leaves at once, with no draft step. Attachments in a thread come with a source the agent can file in a project without passing the file through the conversation.
What you can ask it
"Read my latest Telegram conversations and list who is waiting on me." The agent lists recent threads, opens the ones that ended on someone else's message and returns the people with a line each. Because the list is flat, it decides what counts as waiting from the messages, not from a folder.
"Summarize this thread with the supplier group." The agent finds the group in the list, reads its messages and writes the summary. Nothing is sent.
"Reply to Sam: the invoice goes out Friday." The agent shows you the text and sends it into Sam's thread once you agree. The message appears as coming from you.
Telegram pairs with WhatsApp for teams that run customer chat on both, and with LinkedIn when a conversation started there.
What the Telegram connector can't do
- It does not use the Bot API, so it does not run a bot or post as one.
- It does not sort threads into folders or inboxes. The list is flat.
- It sends text only. Attachments can be read but not sent.
- It does not create groups or channels, change members or settings, or delete messages.
- A send leaves at once. There is no draft or scheduling step.
- It acts as your own account, so it is not a shared inbox for a team.
- The older send operation on the read tool is deprecated and served until 10 December 2026. Use the send tool.
Why through Tulina
The connection was never the hard part.
Most systems now publish their own way for a model to reach them, and that is good: it is what Tulina connects to. The hard part is deciding who may use it, what it is allowed to return, and what the model knows before it asks.
The keys stay in one place
YouTeamOrganisation
Connect Telegram once. The credential is encrypted in a vault, never held in clear, and never pasted into a chat or a workflow file. A person uses their own key where they have one, their team's where they do not, the organisation's after that, so nobody is blocked and nobody holds more access than they need. Revoke it once and everything running on it stops.
What leaves is filtered before it leaves
fields droppedmasked or hashedpseudonymised
Roles decide who can reach Telegram at all. Beyond that, a policy on the connector itself decides what comes back. The data is minimised at the source rather than after it has already reached a model.
Every call is on the record
whatwhowhen
Runs are logged for the organisation. That is what makes an automated action reviewable after the fact, and it is what a security review asks for first.
The run starts knowing your company
namingkey accountsdecisions
A direct connection gives a model access. It does not tell it how your company names things, which accounts matter, or what was decided last quarter. Tulina carries that across every tool, so the same request against Telegram returns something usable instead of something generic.
Processes that run on Telegram
None published yet.