ownRL
Every time you use an AI coding assistant, you correct it.
Core trick
Uses your own test suite to score coding-agent fixes to broken TypeScript files drawn from your past development session traces.
Description
Every time you use an AI coding assistant, you correct it. You rephrase, redirect, steer it back. Those corrections encode what you mean when you say things, your taste, your design directions. Right now that signal disappears. The next session starts from scratch.
We built the loop that captures and refines models with it.
This is the first partial implementation of OpenEnv RFC 005: an RL gym that trains a coding agent on your own dev sessions, scored by your own test suite, so the model gets incrementally better at understanding you specifically.
README excerpt
--- title: MOA RL Environment emoji: 🤖 colorFrom: blue colorTo: purple sdk: docker app_port: 8000 pinned: false --- # MOA RL Environment An OpenEnv-compatible RL environment for training agents on real TypeScript engineering tasks derived from MOA developer session traces. ## API - `POST /reset` — get a new task (broken file + description + tests) - `POST /step` — submit a fix, get vitest reward score - `GET /state` — current environment state