RunBuoy
  • English

      • RunBuoy

      Long-running progress,always within reach

      Keep commands, builds, experiments, and AI agent runs on Mac and Linux visible in Live Activities, on the Lock Screen, and in Dynamic Island.

      Install with an AgentManual installApp Store · Coming Soon

      Built for work that keeps running

      The important status, at a glance

      Bring phases, progress, and results within reach while keeping local data boundaries intact.

      Status that keeps updating

      See each phase, honest progress, and final results in Live Activities, on the Lock Screen, and in Dynamic Island.

      Privacy-first, yours to control

      Core run data stays on your machine by default. The open-source stack is auditable and ready to self-host.

      Native AI agent workflows

      Follow phases, progress, and results from Codex and other agents through the CLI and $runbuoy skill.

      Recommended setup

      Let your agent handle installation

      Copy one prompt to install the RunBuoy Skill and CLI, then verify the machine.

      Recommended

      Install the Skill + CLI

      Your agent checks the environment first and follows the safety rules in the RunBuoy Skill. It pauses for approval before system-level changes.

      • Skill and CLI together
      • Verified after install
      • Approval before system changes
      Complete installation prompt
      Help me install and verify RunBuoy:

1. Use your native Skill installation mechanism to install the RunBuoy Skill from
https://github.com/TANG617/RunBuoy/tree/main/skills/runbuoy
Preserve SKILL.md, agents/openai.yaml, and the references/ directory in full. Do not guess an installation path.

2. After installation, read references/installation.md from that Skill and follow its safety rules to install the RunBuoy CLI. First check command -v runbuoy. If it is missing and uv is available, run:
uv tool install --python 3.12 runbuoy

3. If uv or tmux is missing, or sudo, a system package manager, or a curl installer is required, explain the exact command and wait for my approval.

4. Finally run:
runbuoy --version
runbuoy doctor --json
runbuoy capabilities --json

Only report whether the Skill is discoverable as $runbuoy, the CLI version, local_ready, and delivery status. Do not start pairing, a demo, or any monitored command, and do not upload logs.

      Alternative

      Or install the CLI manually

      Use the public, executable uv flow when you want to control each command yourself.

      1. Prepare dependenciesMake sure tmux and uv are installed on the system.
      2. Install the CLIUse an isolated Python 3.12 tool environment for RunBuoy.
      3. Verify the machineCheck the version, local_ready, and delivery status.
      macOS / Linux · Python 3.12
      uv tool install --python 3.12 runbuoy
runbuoy --version
runbuoy doctor --json
runbuoy capabilities --json

      The agent prompt and manual installation are available now. The App Store release is still being prepared. Check current availability

      Read the complete setup and usage guide

      Data boundaries

      Send status, not the work itself

      Your computer stays in control. RunBuoy presents carefully scoped phases, progress, timing, and results on iPhone.

      Mac / Linux

      Commands and full logs stay local

      RunBuoy Server

      Relays carefully scoped run status

      iPhone

      Live Activities, Lock Screen, Dynamic Island

      • Commands, source, environment variables, and full logs stay on the computer by default.
      • The open-source components are auditable and the Server is ready to self-host.
      Learn about privacy and security boundaries

      RunBuoy

      Keep your next long run visible

      Install the Skill and CLI first. Local Runs remain fully available before you pair an iPhone.

      Install with an AgentManual install