altk-evolve
by AgentToolkit·★ 84·Score 46
Evolve provides AI agents with memory and learning capabilities through MCP server integration.
Overview
Evolve is a system that helps AI agents improve over time by learning from their interactions and experiences. It uses an MCP server for tool integration, vector storage for memory management, and LLM-based conflict resolution to refine its knowledge base. The system maintains a feedback loop, taking conversation traces from agents, extracting key insights into a database, and feeding that information back to improve future performance. It supports both lite mode for existing agent assistants and a full MCP server implementation with comprehensive entity management capabilities.
Try asking AI
After installing, here are 5 things you can ask your AI assistant:
When to choose this
Choose Evolve when you need AI agents that learn and improve from past interactions, particularly for coding or complex multi-step tasks where agents would otherwise repeat the same mistakes.
When NOT to choose this
Avoid Evolve if you need a simple, lightweight solution without persistent memory or if you're not using Python-based agents, as it requires Python 3.12+ and has multiple dependencies.
Tools this server exposes
7 tools extracted from the READMEget_entitiesget_entities(task: str, entity_type: str = "guideline", include_public: bool = False)Get relevant entities for a specific task
get_guidelinesget_guidelines(task: str)Get relevant guidelines for a specific task
save_trajectorysave_trajectory(trajectory_data: str, task_id: str | None, owner_id: str | None)Save a conversation trajectory and generate new guidelines
create_entitycreate_entity(content: str, entity_type: str, metadata: str | None, enable_conflict_resolution: bool, owner_id: str | None, visibility: str = "private")Create a new entity with optional conflict resolution
publish_entitypublish_entity(entity_id: str, user_id: str | None)Make an entity publicly visible to all namespaces
unpublish_entityunpublish_entity(entity_id: str, user_id: str | None = None)Revert an entity to private visibility
delete_entitydelete_entity(entity_id: str)Delete a specific entity by its ID
Comparable tools
Installation
Installation
Prerequisites:
- Python 3.12 or higher
uv(recommended) orpip
From Source:
git clone https://github.com/agenttoolkit/altk-evolve.git
cd altk-evolve
uv venv --python=3.12 && source .venv/bin/activate
uv sync
# Build the UI
cd frontend/ui
npm ci && npm run build
cd ../..From PyPI:
pip install altk-evolveConfiguration
For direct OpenAI usage:
export OPENAI_API_KEY=sk-...Running Services
Start the Web UI and MCP server:
uv run evolve-mcpThe Web UI can be accessed from: http://127.0.0.1:8000/ui/
Claude Desktop Integration
Add to Claude Desktop configuration:
{
"mcpServers": {
"evolve": {
"command": "uv",
"args": ["run", "evolve-mcp"]
}
}
}FAQ
- What is the difference between Lite mode and full MCP server mode?
- Lite mode is designed to integrate with existing agent assistants like Claude Code and Codex without the full Interaction layer. Full MCP server mode includes all features including comprehensive entity management, conflict resolution, and sharing capabilities through the MCP protocol.
- How does Evolve handle conflicting guidelines?
- Evolve uses LLM-based conflict resolution to refine its knowledge base. When new information contradicts existing guidelines, it automatically updates them based on the new insights, improving the knowledge base over time.
Compare altk-evolve with
Last updated · Auto-generated from public README + GitHub signals.