A working reference implementation of a human-led AI software-development organism.
PyAutoScientist is the live system that develops all of the PyAutoLabs software. Plain-English intent becomes tested, released code through an organism of repositories — a set of AI agents that plan, implement, test and release the work, with a human directing it and checkpointing every decision that matters. It is not a framework you install; it is documented so you can fork it and lead your own.
📖 Docs: https://pyautoscientist.readthedocs.io 🍴 Adoption guide: https://pyautoscientist.readthedocs.io/en/latest/adoption/guide.html
This repo is a home for the PyAutoScientist docs and links for now. It will grow as the organism does.
| Organ | Repo | Role |
|---|---|---|
| Mind | PyAutoMind | Where you lead — every piece of work starts here as a plain-English markdown file saying what to do. |
| Brain | PyAutoBrain | The reasoning layer that works out how — classifying, planning and routing each task through specialist agents. |
| Hands | PyAutoBuild | The executor that packages, tags and releases the libraries to PyPI, nightly. |
| Heart | PyAutoHeart | The health monitor whose GREEN/YELLOW/RED verdict is the authoritative "is it safe to release?" gate. |
| Memory | PyAutoMemory | Long-term scientific knowledge — cross-linked literature wikis the agents consult. |
| Gut | PyAutoGut | The storage mirror of Memory — holds condemned material (stale branches, dead code) as recoverable git refs, then voids it on a sweep. |
| Nerves | PyAutoConf | The configuration and serialization layer (autoconf) connecting the organism's conventions to every library. |
The software the organism develops lives across the PyAutoLabs organisation — see the front door for the full map.
Released under the MIT License. Copyright (c) 2026 Jammy2211.