Get started
Kinex documentation
A general harness for arbitrary robots with coding agents — skills, tools and devices on top of the Pi coding agent.
Kinex is a general harness for arbitrary robots with coding agents, built on the Pi coding agent. It gives an agent three layers to work with:
| Layer | Gives the agent | Example |
|---|---|---|
| Skills | Task procedures the agent loads on demand | setup-device, calibrate-camera-extrinsics |
| Tools | Executable operations over devices and computation | device, exec, servo, locate_object_3d |
| Devices | Typed access to hardware, simulators and model services | arm, wrist_cam, world, sam3 |
Kinex lets you work with real hardware the way you already work with a coding agent. Describe the task in plain language and the robot carries it out.
Install and run
curl -fsSL https://raw.githubusercontent.com/stacx-lab/kinex/main/install.sh | bash
cd /path/to/robot-workspace
kinexThen /login to connect a model provider and /model to choose a model.
| Need | For |
|---|---|
| macOS or Ubuntu | Installing and running Kinex |
| Node.js ≥ 22.19 + npm | Kinex itself (the installer can install them) |
curl | The installer |
| Python ≥ 3.11 + uv | Python tools, drivers, exec |
ffmpeg | Video encoding during recording |
Already use Pi, Claude Code or Codex? Kinex plugins currently support these three harnesses: see Plugin systems.
Pick a path
Understand the design of Kinex
- Architecture — skills, tools, devices and who owns what.
- Devices — types, profiles, drivers, frames and clocks.
- Tools & skills — the built-in tools and skills, and how custom ones are discovered.
- Runtime — what Kinex handles for you: devices, discovery, permissions, frames, time and context.
Go further
| Task | Page |
|---|---|
| Choose a model and approval mode; manage conversations, compaction and sessions | Working with the agent |
| Set up new hardware and calibrate it | Onboarding & calibration |
| Run Python, add your own tools and skills | Custom tools & skills |
| Run VLA policies and feedback loops | Using VLAs as tools |
| Find and locate objects with vision models | Using vision models as tools |
| Record what the robot sees and does | Recording |
| Install Pi plugins, or use Kinex from other harnesses | Plugin systems |
Reference: CLI, slash commands, troubleshooting.
About this version
These pages describe Kinex 1.0.0, bundling the Pi coding agent 1.2.0 (@earendil-works/pi-coding-agent). No separate Pi install is needed. Install on macOS and Ubuntu. Node.js ≥ 22.19 is required.
The documentation tracks source commit afeddc6 (2026-10-14). Source references are recorded in each page’s metadata. Kinex and its core/setup packages use the Apache License, Version 2.0. Bundled third-party notices are listed in THIRD_PARTY_NOTICES.txt.