> ## Documentation Index
> Fetch the complete documentation index at: https://docs.remoteagent.chat/llms.txt
> Use this file to discover all available pages before exploring further.

# What is RemoteAgent.CHAT?

> The managed Telegram bridge for any AI coding agent. Control Claude Code, Gemini CLI, Aider, OpenCode, and custom AI tools from your smartphone.

RemoteAgent.CHAT is a lightweight relay that connects any AI coding agent running on your machine to a Telegram bot you control. Once paired, you can send natural-language prompts from anywhere — your phone, another computer, anywhere Telegram runs — and get streaming output back in real time.

There is no cloud IDE, no remote execution environment, and no server you need to manage. Your agent runs on your own machine, with your own API keys, in your own project directory. RemoteAgent.CHAT is purely a relay: it never sees your source code, your prompts in full context, or your API credentials.

The setup takes under five minutes. Download the binary, get a pairing code from [remoteagent.chat/dashboard](https://remoteagent.chat/dashboard), run `remoteagent init` in your project directory, and paste the code when prompted. You're done.

<CardGroup cols={2}>
  <Card title="Any AI runner" icon="robot">
    Works with Claude Code (SDK or CLI), Gemini CLI, Aider, OpenCode, OpenClaw, or any
    custom binary that reads stdin and writes stdout. Switch runners at any time
    with a single flag.
  </Card>

  <Card title="Telegram interface" icon="paper-plane">
    No custom app to install. Use the Telegram client you already have. Commands
    stream back as messages in real time, split automatically when they exceed
    Telegram's 4096-character limit.
  </Card>

  <Card title="Zero trust relay" icon="shield-check">
    Your API key never leaves your machine. RemoteAgent.CHAT routes messages between
    Telegram and your local agent — it does not store prompts, code diffs, or
    credentials at any point.
  </Card>

  <Card title="5-minute setup" icon="bolt">
    One command to install, one init wizard, one pairing code. No Docker, no
    configuration files to write by hand, no infrastructure to provision.
  </Card>
</CardGroup>

## How it works

1. You download the `remoteagent` binary on the machine where your AI agent will run.
2. You log in to [remoteagent.chat/dashboard](https://remoteagent.chat/dashboard), click **Pair new agent**, and get a short pairing code.
3. You run `remoteagent init` in your project directory, choose your AI runner, and paste the pairing code when prompted. The agent is now linked to your Telegram bot.
4. From that point on, any message you send to the bot is forwarded to your agent. Output streams back to Telegram as it is produced.

The relay is event-driven — no polling, no persistent connections from the server side. See [Architecture](/under-the-hood/architecture) for a detailed breakdown.

## Compatibility

<Note>
  The RemoteAgent.CHAT binary is self-contained — no Node.js or runtime required.
  It runs on Linux, macOS, and Windows (WSL recommended on Windows). The runner
  you choose may have its own prerequisites — see the
  [Runners](/runners/overview) section for details.
</Note>