Untether

Name: Untether
Author: Little Bear Apps

Code from anywhere with Claude Code, Codex CLI, Gemini CLI, OpenCode, Pi, and Amp

Telegram Released

Get Started View on GitHub

6:40

Edit

Chats

All Untether 3

littlebearapps.com 6:40 PM

Untether Bot VPS

starting · claude code · 0s 🟧 dir: lba-web

@ 2

Brand Copilot 6:37 PM

Untether Bot VPS

working · claude code · 2m 28s · step 24 ✓ for file in...

Outlook Assistant 6:34 PM

Untether Bot VPS

working · codex · 6m 17s · step 124 ✓ grep -rn...

PitchDocs 6:28 PM

Untether Bot VPS

finished · claude code · 4m 52s · step 67

Scout 6:08 PM

Untether Bot VPS

working · opencode · 32m 14s · step 203 ✓ npm run...

Viewpo (mac) 5:55 PM

Untether Bot Local

finished · claude code · 1m 38s · step 12

Platform SDKs 5:22 PM

Untether Bot VPS

finished · claude code · 8m 45s · step 89

viewpo.io Wed

Untether Bot VPS

cancelled · claude code · 5m 07s · step 12

Contacts

Calls

Chats

Settings

Telegram bridge for AI coding agents

Voice, text, files, and approvals — all from Telegram, wherever you are.

What Untether includes

Every engine gets the core features. Some support deeper integration.

Feature	Claude Code	Codex CLI	OpenCode	Pi	Gemini CLI	Amp
Progress streaming
Session resume
Model override						¹
Model in footer		—	—	—		—
Approval mode in footer		⁴	—	—	²	—
Voice input
Verbose progress
Error hints
Preamble injection
Cost tracking		³		³	³	³
Interactive permissions		—	—	—	—	—
Approval policy		⁴	—	—	²	—
Plan mode		—	—	—	—	—
Ask mode (option buttons)		—	—	—	—	—
Diff preview		—	—	—	—	—
Auto-approve safe tools		—	—	—	—	—
Progressive cooldown		—	—	—	—	—
Subscription usage		—	—	—	—	—
Reasoning/effort levels			—	—	—	—
Device re-auth (/auth)	—		—	—	—	—
Context compaction	—	—	—		—	—
Cross-env resume (/continue)				⁵		⁶

¹ Amp model override maps to --mode (deep/free/rush/smart)

² Toggle via /config between read-only (default), edit files (--approval-mode=auto_edit, files OK but no shell), and full access (--approval-mode=yolo); pre-run policy, not interactive mid-run approval

³ Token usage counts only — no USD cost reporting

⁴ Toggle via /config between full auto (default) and safe (--ask-for-approval=untrusted, untrusted tools blocked); pre-run policy, not interactive mid-run approval

⁵ Pi requires provider = "openai-codex" in engine config for OAuth subscriptions in headless mode

⁶ Amp requires an explicit thread ID; no "most recent" mode

Requires a AI Coding Agents account

Get started in 60 seconds

Copy this prompt and paste it into any AI assistant:

I want to install Untether (https://github.com/littlebearapps/untether) to control Claude Code from Telegram. Please guide me step-by-step, including creating the Telegram bot and running the setup wizard.

Or configure it yourself

Install via command line

See our user guides

"I was walking Normi and had an idea for a fix. By the time I got home and opened my laptop, I'd forgotten half of it."

—

Nathan

More about Untether

Terminal listing MCP server processes across Claude Code sessions — each session spawns about 14 servers, consuming over 100 MB each

8 March 2026 · 18 min read ·

dogfooding testing

Dogfooding found 22 bugs my 1,548 tests missed

Dogfooding Untether and PitchDocs daily — and Outlook Assistant when the task calls for it — found 22 bugs that 1,548 automated tests missed. The bugs that matter live in the transitions between states — sleeping and waking, busy and stuck, present and away.

See all blog posts

Help & support

v0.34.4

Getting Started

Set up Untether step by step

How-To Guides

30 practical guides

Tutorials

6 step-by-step tutorials

All Untether Help

Guides, reference, and more

Questions about Untether

What is Untether?

Untether is a Telegram bridge that connects AI coding agents to your phone. Send tasks by text or voice, watch progress in real time, and approve changes with inline buttons. It works with Claude Code, Codex CLI, OpenCode, Pi, Gemini CLI, and Amp.

Does Untether send my code to Little Bear Apps?

No. All communication flows between Telegram and the agent running on your machine. Little Bear Apps never sees your code, messages, or agent output.

Do I need a Telegram account?

Yes. You need a Telegram account and a bot token from @BotFather. The bot runs on your machine and only responds to your Telegram user ID.

Can I use voice notes?

Yes. Untether transcribes voice notes via a configurable Whisper-compatible endpoint, and the transcription is sent to your agent as a text message.

Does it work with other AI coding agents?

Yes. Untether supports Claude Code, Codex CLI, OpenCode, Pi, Gemini CLI, and Amp out of the box. The plugin system allows adding custom agent integrations.

Can I control multiple projects at once?

Yes. Register multiple repos with untether init and switch between them from Telegram using project-specific commands or forum topics.

Is there a cost for using Untether?

Untether itself is free and open source (MIT licence). You pay for your AI agent's API usage as normal. The Telegram Bot API is free.

How do I restrict access to my bot?

Set allowed_user_ids in untether.toml to your Telegram user ID. The bot ignores messages from anyone else.

How do I set up Untether?

Run untether in your terminal — the setup wizard creates a Telegram bot, picks your workflow, and connects your chat. You can also pass an existing bot token with --bot-token.

Can I continue a previous session?

Yes. Reply to any bot message to resume that session, or the agent picks up where it left off automatically. You can also use /continue to pick up a session you started in your terminal — works with Claude Code, Codex, OpenCode, Pi, and Gemini.

Can I run multiple tasks at the same time?

Yes. Use forum topics or separate project chats to run tasks in parallel. Branch targeting with @branch runs in isolated git worktrees so tasks don't interfere with each other.

Can I automate recurring tasks?

Yes. Untether supports cron expressions and webhook triggers in the config file. You can also use Telegram's built-in message scheduling for simpler automation.

Roadmap

Have a feature idea? Let us know!

99%

v0.35.0

Largest release since initial publication — 147 issues across 14 release candidates. Three workflow modes (assistant/workspace/handoff), cross-environment resume (/continue), /config UX overhaul, agent-initiated file delivery (outbox), interactive plan outline with hold-open approval, Codex/Gemini approval policies, auto-continue for Claude Code, stall diagnostics with CPU-aware suppression, orphan progress cleanup, cost budget toggles, model metadata for all 6 engines, and 60+ bug fixes.

TBD

149/150 issues