altk-evolve
Active·★ 85·Apache-2.0·Updated 2026-05-29
★ Trending★ Workflow Automation★ Memory & Context
Self improving agents through iterations
Evolve is a system designed to help agents improve over time by learning from their trajectories. It uses an MCP server for tool integration, vector storage for memory, and LLM-based conflict resolution to refine its knowledge base.
#agent-learning#agent-memory#agentic-workflows#ai-agents#claude-code#claude-code-plugin#claude-code-plugins#codex
01
Features
01MCP Server: Exposes tools to get guidelines and save trajectories.
02Conflict Resolution: Intelligently merges new insights with existing guidelines using LLMs.
03Trajectory Analysis: Automatically analyzes agent trajectories to generate guidelines and best practices.
04Milvus Integration: Uses Milvus (or Milvus Lite) for efficient vector storage and retrieval.
02
Compatibility
Python
Python 3.12+
Verified via docs
MCP
MCP Client
Verified via docs
Milvus
Milvus/Milvus Lite
Verified via docs
03
Quick start
1
$ git clone https://github.com/AgentToolkit/altk-evolve.git
2
$ cd altk-evolve
3
$ uv venv --python=3.12
4
$ source .venv/bin/activate
5
$ uv sync
04
Use cases
↳Continuous improvement of AI agents in production
↳Learning from agent trajectories to refine guidelines
↳Integrating with MCP-compatible tools and clients
05
Alternatives
letta★ 23.0k
Letta is the platform for building stateful agents: open AI with advanced memory that can learn and self-improve over time.
DesktopCommanderMCP★ 6.1k
This is MCP server for Claude that gives it terminal control, file system search and diff file editing capabilities
Auto-claude-code-research-in-sleep★ 11.0k
ARIS ⚔️ (Auto-Research-In-Sleep) — Claude Code skills for autonomous ML research: cross-model review loops, idea discovery, and experiment automation via Codex MCP
agents-best-practices★ 1.1k
Provider-neutral Agent Skill for Codex, Claude Code, and agentic harness design.
awesome-claude★ 250
HeyClaude is a curated registry and distribution surface for Claude and AI-workflow assets: agents, MCP servers, skills, commands, hooks, rules, guides, tools, jobs, Raycast feeds, static data exports, and an npm MCP package.
Related searches
Comments
Log in to leave a comment
- TTaylor JohnsonApr 19, 2026
Good for benchmarking how much agents improve across task domains over iterations
- SScout DavisApr 9, 2026
Self-improving agents through iteration is the right approach to capability growth
- CCameron ClarkMar 9, 2026
The evolution framework avoids the catastrophic forgetting issues common in naive approaches