A lightweight Python library that allows any LLM agent to self-improve through feedback, without retraining models.

How to effectively finetune CV/LLM models (without local gpu)

Self improving agents through iterations

self-replicating ai agent built for you

AI agent pipeline orchestration for Claude Code. Deploy specialized agents in coordinated pipelines with self-learning capabilities, mesh coordination, Raft consensus, CRDT knowledge graphs, and pipeline debugging. 15 production-grade skills with reasoning protocols, 20+ agent types, 15 ready-to-run templates. Zero infrastructure.

The Self-Learning AI Framework is an autonomous system that analyzes, modifies, and optimizes its own code. It supports natural language interaction, framework creation, code suggestions, PDF-based learning, and package management. Ideal for automating code reviews, agent creation, and AI-driven development.

A set of pi-extensions, including pi-continuous-learning that continually improves you pi coding agent experience

CodeGraph MCP - Give persistent learning to your coding agent for a codebase across teams & token saving superpower to the coding agent

LLM Powered, Autonomous, Self-Learning, Self-Improving, Research System

This project implements an AI that teaches itself to play the game of Nim using Q-learning, a form of reinforcement learning. By playing games against itself, the AI learns optimal strategies for playing Nim, eventually improving its performance by updating a reward table based on game outcomes.

Reasona is an AI system that improves itself. It answers questions using your documents and corrects wrong answers automatically for better future results.

Chess engine implemented with MCTS, neural network and multiprocessing in self-play to generate training data

Discover, evaluate, and manage AI skills from community sources. A security-aware framework for continuous skill acquisition.

Enable the Agent to gain experience from tasks just like humans do.