Any AI coding agent.
One Telegram bot.
Pick your preferred AI runner.
Connect Telegram in under 5 minutes.
Ship code from anywhere.
Three steps.
Then just type.
We obsessed over the onboarding so you don't have to.
One curl command. The wizard helps you pick and install your AI runner — it even installs it for you if it's not there yet. Your API keys never leave your machine.
Log in to remoteagent.chat — your pairing code is waiting in the dashboard. Paste it into the remoteagent init wizard and your Telegram is linked instantly.
Message your bot in natural language. "Add input validation to the signup form." "Fix the 500 error in /api/users." "Push to main."
just type, your agent does the restEverything your AI agent can do.
From your pocket.
No BotFather, no webhook setup. Run remoteagent init in any repo and it connects automatically. Switch between projects with a single message.
Output streams directly to your Telegram chat as it's produced. Watch your agent work in real time — file edits, test runs, git output, errors and all.
Context is preserved across messages. Start a refactor on your desktop, then continue from your phone when you leave — your agent is always ready.
Your API key never leaves your machine. Commands route through Redis — we never see your credentials, your code, or your prompts.
Send a Telegram voice message and it gets transcribed instantly. Drop a screenshot of a bug and your agent sees it too.
Your agent generates a diff, a report, or a chart? It lands directly in your Telegram chat. Files up to 50 MB are delivered instantly — no SSH, no manual download.
Pick your AI runner.
We handle the rest.
RemoteAgent.CHAT works with the tools you already use — or helps you install a new one in seconds.
No lock-in, ever.
Runs the claude CLI binary as a subprocess. Ideal if you already use Claude Code and want to keep your existing setup. Requires an Anthropic API key.
In-process Anthropic SDK with full tool use, streaming, permission control, and conversation history. The most capable option — bundled with RemoteAgent.CHAT, zero extra install. Requires an Anthropic API key.
Google's open-source coding agent. Fast, capable, with a free tier. A solid alternative if you want to use Gemini models instead of Claude.
Open-source terminal AI coding agent by the SST team. Supports 75+ providers — including OpenCode Zen (pay-as-you-go, 37+ models, no provider key needed). Runs non-interactively via opencode run, with full filesystem access and no lock-in.
OpenAI's terminal coding agent. Runs codex exec --json non-interactively with structured output. Supports API key auth or device login from any browser — no browser required on the server.
The battle-tested open-source coding assistant. Works with any LLM — OpenAI, Anthropic, Ollama, and more. Highly configurable, loved by the community.
Self-hosted AI gateway. Supports Claude, OpenAI, and local models — your data never leaves your machine.
⚠️ Beta — not fully tested.
Using something else? RemoteAgent.CHAT supports any binary via the custom runner.
The control plane
for your AI agent.
Your AI tool handles the code.
RemoteAgent.CHAT handles everything else — daemon, Telegram, notifications, multi-project.
remoteagent init in every repo — each one becomes an independent, always-on agent in minutes. remoteagent list to see them all. start and stop to manage the lifecycle. Switch from your phone with /repo — no config to duplicate, no process to babysit.
No terminal to open, no session to start. While your agent is running, commands from Telegram are dispatched instantly — whether you send them at 2pm or 2am.
Send a task, lock your phone. Get a Telegram notification when it's done. You don't watch the terminal — you review the result.
Full session history, agent status monitoring. See what ran, when, and for how long — from any browser.
RemoteAgent.CHAT isn't a replacement for your AI coding tool — it's the layer on top. Keep using whichever agent you love. We give it a Telegram interface, a dashboard, and a background process that runs while you work.
Try everything free.
For 14 days.
Full Pro access from day one. No credit card. No feature limits. If it doesn't change how you work, cancel in one click.
- ✓ Unlimited projects
- ✓ Agent offline alerts
- ✓ Voice input — unlimited
- ✓ Unlimited seats
- ✓ Custom integrations
- ✓ Dedicated support
- ✓ SLA & invoicing
Bring your own AI. RemoteAgent.CHAT is pure relay infrastructure — we never access your API keys or your codebase. Your credentials stay on your machine. Prompts and outputs are relayed through our servers and optionally saved as session logs for your own review. Every user connects directly to their AI provider of choice.
14 days free.
Full Pro. No card.
Install in 3 commands. If it doesn't ship faster, cancel — no questions asked.
Frequently asked questions
What is RemoteAgent.CHAT?
RemoteAgent.CHAT is a Telegram bridge for AI coding agents. It lets you send prompts to Claude, Gemini, Aider, OpenCode or any custom AI runner from your phone — and stream the output back in real time, wherever you are.
Do you store my API keys?
No. Your API key lives only on your machine in ~/.remoteagent/agents/ and is never transmitted to our servers. The agent calls your AI provider directly from your own machine.
Which AI runners are supported?
Claude Code SDK, Claude Code CLI, Gemini CLI, Aider, OpenCode (including Zen), OpenClaw, and any custom binary via stdin/stdout. More runners can be added via the custom runner type.
How long does setup take?
Under 5 minutes. Install the CLI with one curl command, run remoteagent init, paste the pairing code from the dashboard, and you're live. No server config, no ports, no reverse proxy.
Does it work on any server or OS?
Yes. RemoteAgent.CHAT runs on Linux (x64 and arm64), macOS (Intel and Apple Silicon), and Windows. Any machine that can run your AI agent can run RemoteAgent.CHAT — VPS, home server, cloud VM, even a Raspberry Pi.
What happens if my server goes offline?
The bot will alert you as soon as the agent goes offline. Your projects and session history are safe on our end. When the server comes back, the agent reconnects automatically and you can resume immediately.
Can I have multiple projects?
Solo supports 1 project. Pro supports unlimited. If you upgrade to Pro, add more projects, then downgrade back to Solo, your other projects are kept in the database — only 1 stays active. You can switch which one is active from the dashboard at any time.
Is the trial really free, no card required?
Yes. 14 days of full Pro access — no credit card, no commitment. You only need to add a payment method when you decide to subscribe after the trial ends.
Can I cancel or delete my account anytime?
Yes. You can cancel your subscription or permanently delete your account at any time from the dashboard or via the bot — no questions asked, no waiting period.
Is my code or conversation sent to RemoteAgent.CHAT servers?
Only the raw text of your prompts and agent outputs passes through our relay to reach Telegram. We never see your codebase, your file system, or your AI provider credentials. Session logs are optional and only visible to you.