Stablehand
The self-hosted agent runtime for companies that own their learning loop
Early access · TypeScript · MIT
Open source
We open-source the tools we run our own AI work on: the runtime our agents live in, the layer that publishes their work, the playbook our coding agents follow and the context graph behind our daily briefings. Each one runs on your machine or in your own cloud, under the MIT licence.
Artifact Use
The open publishing layer for agent work: stable links, access gates and review that flows back to the agent
TypeScript · MIT
AI-native Claude plugins
The working playbook of an AI-native engineering team, packaged as plugins for Claude Code and Codex
MIT
WhatsApp context CLI
Your WhatsApp history as a private context graph that you and your agents can query
Python and Go · MIT
iofold auto-evals
Evals written from your team’s ratings, checked against your reviewers, and used to promote better agent versions
TypeScript · MIT
Work with us
Tell us what the work is, what data it touches and what constraints apply. We’ll work out what to build and how to measure it.