runner

Where terminal agents work together. Claude Code, Codex, Copilot CLI and pi on the same task, in one mission, each keeping its own TUI in a real terminal.

What is runner?

Where terminal agents work together. Claude Code, Codex, Copilot CLI and pi on the same task, in one mission, each keeping its own TUI in a real terminal.

Overview of runner

What it is

Key features

Use cases

Getting started

License and pricing

What it is

Runner is a desktop application that brings multiple CLI coding agents together in reusable crews. Each agent keeps its own terminal user interface (TUI) in a real PTY, and all agents share a mission feed. Sessions and history remain on the user's machine. Runner runs natively on macOS (Apple Silicon) and Windows (x64).

Key features

  • Reusable crews: Compose saved roles into a crew, choose a lead, and define shared conventions. Each slot can be assigned an agent and model; overrides can be adjusted before launching a mission.
  • Missions and coordination: Each mission gives every crew slot a live terminal and a shared feed. Agents coordinate and bring questions to the user. Individual slots can be stopped, resumed, or restarted without affecting others.
  • Real terminals: Each agent runs in its own real PTY on a GPU-drawn alacritty_terminal grid, with mouse reporting, IME input (including Pinyin), file‑path paste and click‑to‑open, and a shell drawer beneath every chat and mission.
  • Split tabs and windows: Tabs can be split arbitrarily, panes can be dragged to reorder, tabs can be grouped into folders, and additional windows can be opened with ⇧⌘N or Ctrl+Shift+N.
  • Desktop notifications: When Runner is in the background, notifications are sent when an agent needs input, finishes a turn, or fails. Clicking a notification opens the relevant chat or mission. Chat notifications are on by default; mission notifications are opt‑in.
  • Background persistence: Agents keep working when the app is closed. Sessions live in runnerd, the background service. Quitting or crashing the app and reopening it restores the same live terminals.

Use cases

  • Coordinating a team of coding agents (e.g., coder, reviewer, QA) on a single task.
  • Running a multi‑agent workflow where agents share a common mission feed and can hand off tasks.
  • Using a saved crew configuration to repeatedly launch the same team of agents on different projects.
  • Driving agent missions programmatically from a script or another agent via the runner CLI.

Getting started

Install and sign in to at least one supported agent CLI (Claude Code, Codex, Antigravity CLI, GitHub Copilot CLI, or pi) before starting Runner. Runner detects the CLI on your PATH. Download Runner from the official website (runnersh.dev). On macOS, the first launch installs the runner command to ~/.local/bin (or a writable /usr/local/bin if already on the login PATH). On Windows, Runner adds it to the user PATH. Example of driving a mission from the terminal: sh mission=$(runner mission start --crew --goal-file - -q < brief.md) runner mission feed $mission --follow runner msg post --mission $mission --to message runner mission answer $mission runner mission stop $mission For a full list of commands, run runner help.

License and pricing

Runner is open source under the MIT license. There is no pricing information stated; it is classified as open source.

Based on the project's public README and repository metadata.

Add this badge to your README

Featured on YunoDB Badge styles & help

Maintain this project? Show that it is featured on YunoDB.

Markdown[![Featured on YunoDB](https://yunodb.com/badge/yicheng47runner.svg)](https://yunodb.com/tools/yicheng47runner?utm_source=github&utm_medium=badge)
HTML<a href="https://yunodb.com/tools/yicheng47runner?utm_source=github&utm_medium=badge"><img src="https://yunodb.com/badge/yicheng47runner.svg" alt="Featured on YunoDB" height="20"></a>