OpenClaw-RL

Open-source asynchronous reinforcement-learning framework for continuously optimizing personal and general AI agents from live next-state signals such as user replies, tool outputs, terminal results, and GUI state changes.