GH Repository · aiming-lab
MetaClaw
🦞 Just talk to your agent — it learns and EVOLVES 🧬.
- stars
- 3,496
- 30-day movement
- +10/day
- Related entries
- 60
- Connections
- 1
agent-appagentfine-tuningreinforcement-learningpythonllmskill-learninglorametaclawai-agentopenclawonline-learningcontinual-learningPythonmeta-learningtinker
MetaClaw is a Python agent app from aiming-lab that lets an agent learn and evolve through conversation. Its topics indicate it covers online learning, fine-tuning (LoRA), reinforcement learning, meta-learning, and skill learning for LLM-based agents.
Use it when you want an agent that continually learns from your interactions rather than staying static.
Use it to
- Fine-tune an agent with LoRA from conversations
- Apply online and continual learning to an LLM agent
- Train agents with reinforcement learning
- Build skill-learning agent workflows
For Developers building adaptive, learning LLM agents
- Role
- agent-app
- Language
- Python
- Licence
- MIT
- Forks
- 455
- Open issues
- 7
- Last push
- 2026-06-07
- Latest release
- v0.1.0 · 2026-03-10
topicsagentonline-learningfine-tuninglorareinforcement-learningmeta-learning