Untether Help

Untether has three workflow modes that control how conversations continue and how sessions are organised. Each mode suits a different working style. · 3 min read

Bind a chat or forum topic to a specific project and branch, so every message runs in the right directory automatically — no need to prefix with /<project> e... · 2 min read

Running agents remotely means they can rack up costs while you're not watching. Untether tracks API costs per run and per day, with configurable budget limit... · 4 min read

Resume a coding session you started in your terminal — from Telegram, while away from your desk. · 3 min read

Set up Untether for local development, run checks, and test changes safely via the dev instance before releasing to production. · 3 min read

Untether records session events as they stream from the agent, so you can export a full transcript of any run directly from Telegram — review what your agent... · 2 min read

Send files to your repo or pull results back — directly from Telegram on any device. Useful when you need to share a spec, screenshot, or config file with yo... · 4 min read

Untether works in Telegram group chats, letting multiple people interact with coding agents from any device. This guide covers adding the bot to a group, res... · 3 min read

Adjust Untether's behaviour without editing config files or restarting — tap buttons right in Telegram. The /config command opens an interactive settings men... · 5 min read

When Claude Code runs in permission mode, Untether shows inline buttons in Telegram so you can approve or deny tool calls from your phone. · 5 min read

Untether lets you override which model the agent uses and its reasoning level, per chat or per engine — all from Telegram, without editing config files or re... · 2 min read

Untether runs as a long-lived process, typically in a terminal or managed by a process supervisor (systemd on Linux, etc.). This guide covers health checks, ... · 7 min read

When you're away from the terminal, you need confidence that your agent won't go off-script. Plan mode controls how Claude Code handles permission requests t... · 5 min read

Untether injects a context preamble at the start of every agent prompt, telling the engine it's running via Telegram and requesting structured end-of-task su... · 3 min read

Projects let you target any registered repo from Telegram — no need to restart Untether or change directories. Send /myapp fix the tests from your phone and ... · 1 min read

Bind a Telegram chat to a project so messages in that chat automatically route to the right repo. · 1 min read

There are several ways to run tasks on a schedule: the /at command for quick one-shot delays, Telegram's built-in message scheduling, and Untether's trigger ... · 3 min read

Untether gives remote access to coding agents on your server, so locking down who can interact with the bot and what files they can access is important. This... · 6 min read

See how much work your agents have done while you've been away. The /stats command shows per-engine session statistics — run counts, action totals, and durat... · 2 min read

Keep tabs on your Claude Code subscription from anywhere — Untether surfaces usage directly in Telegram. This guide covers checking usage on demand and enabl... · 2 min read

Different tasks suit different agents. Switch engines on the fly — use Claude Code for deep refactors, Codex for quick fixes — without restarting or reconfig... · 2 min read

Topics bind Telegram forum threads to a project/branch context. Each topic keeps its own session and default engine, which is ideal for teams or multi-projec... · 3 min read

Common issues and fixes for Untether. If your agent isn't responding, messages aren't arriving, or something looks off — start here. · 16 min read

If Untether is running as a systemd service, stop it first: · 1 min read

Untether publishes releases to PyPI. To upgrade to the latest version: · 2 min read

Untether shows progress messages as the agent works, updating in real time. Control how much detail you see — from compact summaries to full tool details — s... · 2 min read

Dictate coding tasks hands-free — while walking, driving, or away from a keyboard. Untether transcribes your Telegram voice notes via a Whisper-compatible en... · 1 min read

Untether can start agent runs automatically from external events (webhooks) or on a schedule (cron). Both use the same engine pipeline as Telegram messages, ... · 6 min read

Run tasks on multiple branches in parallel without touching your main checkout. Untether creates isolated git worktrees so you can kick off work on @feat/aut... · 1 min read

Untether supports entrypoint-based plugins for engines, transports, and commands. · 2 min read

Untether supports both stateless and chat modes for session handling. In stateless mode, each message starts a new session unless you reply to continue. In c... · 3 min read

This tutorial shows you how to use different engines for different tasks and set up defaults so you don't have to think about it. Swap between Claude Code, C... · 6 min read

Untether can handle follow-up messages in two ways: chat mode (auto-resume) or stateless (reply-to-continue). Both work across all your devices — start a con... · 3 min read

This tutorial shows you how to register repos as projects and run tasks on feature branches — all from Telegram, without touching a terminal. Jump between re... · 5 min read

This page documents Untether’s user-visible command surface: message directives, in-chat commands, and the CLI. · 5 min read

Untether reads configuration from ~/.untether/untether.toml. · 15 min read

This page documents how Untether resolves run context (project, worktree/branch, engine) from messages. For step-by-step usage, see Projects and Worktrees. · 4 min read

Untether runs two isolated instances on lba-1: staging (PyPI/TestPyPI release) and dev (local editable source). They use separate Telegram bots, separate con... · 7 min read

Untether supports a small set of environment variables for logging and runtime behavior. · 3 min read

When an engine fails, Untether scans the error message and shows an actionable recovery hint above the raw error. The raw error is wrapped in a code block fo... · 7 min read

Quick definitions for terms used throughout the Untether documentation. · 5 min read

Structured, repeatable integration test process run against @untether_dev_bot before every release. Tests exercise all 6 engines across the full feature surf... · 20 min read

Untether supports three workflow modes inherited from takopi. Each mode configures three settings that control session continuation and resume line display. · 3 min read

Untether’s public plugin API is exported from: · 5 min read

Community and third-party plugins that extend untether. · 1 min read

Below is the implementation spec for the AMP CLI (Sourcegraph) runner shipped in Untether. · 3 min read

amp -x --stream-json writes one JSON object per line (JSONL) using a Claude Code-compatible protocol with a type field discriminator. · 1 min read

This document describes how the AMP runner translates AMP CLI --stream-json JSONL events into Untether events. · 4 min read

Below is a concrete implementation spec for the Anthropic Claude Code (“claude” CLI / Agent SDK runtime) runner shipped in Untether (v0.3.0). · 14 min read

claude -p --output-format stream-json --input-format stream-json --verbose writes one JSON object per line (JSONL) with a required type field. · 2 min read

This document describes how the Claude Code runner translates Claude Code CLI JSONL events into Untether events. · 6 min read

codex exec --json writes one JSON object per line (JSONL) to stdout. Each line is a top-level thread event with a type field. · 5 min read

This document describes how Codex exec --json events are translated to Untether's normalized event model. · 7 min read

Below is the implementation spec for the Gemini CLI runner shipped in Untether. · 3 min read

gemini -p --output-format stream-json writes one JSON object per line (JSONL) with a required type field. · 1 min read

This document describes how the Gemini runner translates Gemini CLI --output-format stream-json JSONL events into Untether events. · 3 min read

This runner integrates with the OpenCode CLI. Shipped in Untether v0.5.0. · 2 min read

opencode run --format json writes one JSON object per line (JSONL) to stdout. Each line has a type field indicating the event type. · 3 min read

This document describes how OpenCode JSON events are translated to Untether's normalized event model. · 2 min read

Below is a concrete implementation spec for the Pi (pi-coding-agent CLI) runner shipped in Untether (v0.5.0). · 3 min read

pi --print --mode json writes one JSON object per line (JSONL) with a required type field. These are AgentSessionEvent objects from @mariozechner/pi-agent-core. · 1 min read

This document describes how the Pi runner translates Pi CLI --mode json JSONL events into Untether events. · 4 min read

This document is normative. The words MUST, SHOULD, and MAY express requirements. · 16 min read

TelegramClient is the single transport for Telegram writes. It owns a TelegramOutbox that serializes send/edit/delete operations, applies coalescing, and enf... · 8 min read

The trigger system lets external events start agent runs automatically. Webhooks accept HTTP POST requests (GitHub pushes, Slack alerts, PagerDuty incidents)... · 15 min read