longmans

longmans / self-evolve

Public

Self evolve extension for openclaw. Let your claw grow continuously.

16
3
89% credibility
Found Mar 13, 2026 at 16 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

Self-evolve is a plugin for openclaw AI agents that enables self-improvement by retrieving relevant past experiences, learning from user feedback, and updating memory over multiple conversation turns.

How It Works

1
📰 Discover Self Evolve

You hear about a helpful add-on that lets your AI chat buddy learn and improve from everyday talks.

2
📥 Add it to your AI

You grab the add-on and slip it into your AI helper, making it ready to grow smarter.

3
🔗 Link a thinking service

You connect a smart service so your AI can understand chats deeply and start remembering lessons.

4
🚀 Turn learning on

You flip the switch to enable self-improvement, and everything springs to life.

5
💬 Chat and share feedback

You have natural conversations, saying 'great!' for wins or 'not quite' for fixes to guide it.

6
🧠 Watch it recall memories

You see your AI pull up past experiences before replying, making answers sharper and faster.

🎉 AI masters new skills

Over time, your chat buddy evolves, handling tricky tasks better with less effort from you.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 16 to 16 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is self-evolve?

self-evolve is a TypeScript plugin for openclaw that makes your self-hosted AI agent continuously improve by retrieving past episodic memories and injecting them into prompts before responses. It aggregates multi-turn tasks, detects feedback like "great job" or "not working," scores rewards with an LLM like gpt-4o-mini, and updates Q-values while storing new experiences—all to learn skills with fewer tokens than full retraining. Users get a self-evolving agent that handles tools better over time, configurable via CLI like `openclaw config set plugins.entries.self-evolve`.

Why is it gaining traction?

It stands out by blending retrieval-augmented generation with lightweight RL on memories, letting agents evolve on self-hosted setups without heavy fine-tuning or external services. The hook is log-visible progress: see retrieval hits, learning triggers, and modes like "tools_only" to minimize OpenAI costs. Optional remote sharing pools experiences across instances for faster bootstrapping.

Who should use this?

DevOps engineers running openclaw on GitHub self-hosted runners, Docker, or Kubernetes for agentic workflows like bash automation or config troubleshooting. Ideal for teams building self-hosted AI starter kits where agents need to adapt to repeated tasks without manual intervention.

Verdict

Worth testing for self-evolving agents in prototypes—strong docs, one-shot config, and full tests lower the risk despite 16 stars and 0.9% credibility score signaling early maturity. Skip for production until more real-world miles.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.