kayba-ai/agentic-context-engine

🧠 Make your agents learn from experience. Now available as a hosted solution at kayba.ai

66
/ 100
Established

This project helps AI agents learn from their past experiences to avoid repeating mistakes and to improve performance over time. It takes in feedback or observed errors from an agent's task execution and produces refined strategies. This results in AI agents that become more consistent, efficient, and accurate without human intervention.

1,978 stars. Actively maintained with 101 commits in the last 30 days.

Use this if your AI agents are making repetitive errors, forgetting successful approaches, or if you need them to continuously improve their task execution without constant fine-tuning.

Not ideal if you are looking for a simple API wrapper for LLMs without any need for persistent learning or autonomous self-improvement capabilities.

AI-agent-performance autonomous-systems workflow-automation LLM-optimization
No Package No Dependents
Maintenance 22 / 25
Adoption 10 / 25
Maturity 13 / 25
Community 21 / 25

How are scores calculated?

Stars

1,978

Forks

250

Language

Python

License

MIT

Last pushed

Mar 12, 2026

Commits (30d)

101

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/agents/kayba-ai/agentic-context-engine"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.