thith

thith / teleport

Public

Telegram bridge for AI coding agents (Claude Code, Codex, Gemini) — short reports out, short instructions back, no chat mirroring

17
10
80% credibility
Found May 20, 2026 at 17 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
JavaScript
AI Summary

Teleport is a small tool that bridges your AI coding assistants (like Claude Code, Gemini, or Codex) with Telegram, so you can receive short progress updates on your phone and send back quick instructions while you're away from your computer. Instead of mirroring every detail of your AI session to your phone (which would be overwhelming on a small screen), it sends only the important highlights—like a human assistant who pops their head in to say 'finished the first part, should I continue with the next step?' You reply with a simple 'yes' or new directions, and your AI picks up exactly where it left off. It's designed for developers who want to step away from their desk but still stay informed about long-running AI tasks.

How It Works

1
💻 You start working with an AI coding assistant

You're deep in a coding session, building something exciting with an AI helper at your side.

2
🚶 You need to step away from your desk

A meeting, lunch, or errands call you away—but you want to know when your AI assistant finishes important tasks.

3
📱 You connect your AI assistant to Telegram

You set up a simple bridge so your AI can send you messages on your phone, just like texting a helpful colleague.

4
✈️ You leave and ask your AI to keep you posted

Before heading out, you tell your assistant to 'ping me when you're done' or 'text me after each change.'

5
Your phone buzzes with an update
👍
You like what you see

You reply with a quick 'sounds good, keep going' and your AI picks it up and continues working.

🔄
You want to change direction

You reply with new instructions like 'actually, let's try a different approach' and your AI adjusts course.

🎉 Your AI keeps working while you live your life

You stayed informed and in control without being chained to your desk—like having a reliable assistant who keeps you in the loop.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 17 to 17 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is teleport?

Teleport is a JavaScript-based Telegram bridge that lets AI coding agents like Claude Code, Codex, and Gemini send you short progress reports and receive short replies back. Instead of mirroring your entire terminal session to your phone—which gets messy on a small screen—it forwards only the important updates while you step away from your desk. You tell the agent "ping me when done," and it sends a brief message to Telegram and waits for your response if needed.

Why is it gaining traction?

The hook is simplicity: full session mirrors are overwhelming on mobile, but being completely disconnected from your agent means missing critical moments. Teleport splits the difference by acting like a human assistant in the next room—short updates in, short instructions out. It works with multiple AI agents and doesn't require any external dependencies beyond Node.js. The setup process uses a clever auto-configuration prompt that wires Telegram support into all your agents' global config files at once.

Who should use this?

This is for developers who run AI coding agents for long tasks and need to step away from their desk without losing oversight. Remote workers, engineers who pair with agents across different machines, or anyone running overnight builds will find this most useful. If you find yourself constantly checking your terminal to see if your agent finished or made a wrong turn, Teleport replaces that habit with a simple ping.

Verdict

At 17 stars with a credibility score of 0.800000011920929%, Teleport is young and unproven at scale. The documentation is solid and the zero-dependency design means low risk, but there's no test suite mentioned and the project appears maintained by a single developer. Worth trying on a non-critical project first—it's lightweight enough to be worth the experiment if you frequently step away from running agents.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.