---
title: "What is Telegram-Auto?"
locale: en
url: https://tlg.builera.app/en/docs/start/overview
verified: 2026-10-24
app_version: "2.0"
---

# What is Telegram-Auto?

Telegram-Auto helps a team run many personal Telegram accounts from one shared inbox, with a safety gate in front of everything that goes out.


Telegram-Auto is a web app for **a team** to run **many personal Telegram accounts** together: read and answer messages in one shared inbox, manage customers, run controlled campaigns, administer groups and channels, and let AI agents work through the API or MCP.

## What the system is made of

| Part | What it does |
| --- | --- |
| **Web app** | Where you and your teammates work: inbox, customers, campaigns, settings. |
| **Worker nodes** | Servers owned by the workspace (Hetzner, created for you, or your own Windows machine) that hold the Telegram sessions and do the work. |
| **Proxies** | Every account goes out to the Internet through **exactly one proxy of its own**. Without a proxy, the account does not connect. |
| **AI connections** | The door for AI agents and other systems to use REST API v2, MCP and webhooks — with their own rights, never your full rights. |

## The path of one message

1. You (or a campaign, or an AI agent) send a message.
2. It is queued for the right account, in the right order within the conversation.
3. The **safety gate** checks before handing it out: is the account healthy, has Telegram asked it to slow down, is the recipient on the do-not-contact list, is there limit left.
4. The account's worker node takes the job, sends through the account's proxy and reports back.
5. The inbox shows the real state: sent, failed with a reason, or "unconfirmed" when it must be checked against Telegram.

> [!IMPORTANT]
> The safety gate applies the same way to the app, campaigns, automation, the REST API and MCP. Nothing goes around it.

## Two limit modes

Each account runs in **Free** mode (default for new accounts, unlimited unless you set a limit) or **Safe** mode (default limits, warm-up and working hours). In both modes the hard protections stay on. Details in [Limits](https://tlg.builera.app/en/docs/reference/limits).

## Read next

- [Reference](https://tlg.builera.app/en/docs/reference) — REST, MCP, events, error codes, scopes, limits, states.
- [MCP](https://tlg.builera.app/en/docs/reference/mcp) — for AI agents.