Knowledge Graph Memory for AI Agents
Ryumem is an open-source bi-temporal knowledge graph memory system that gives your AI agents intelligent, long-term memory with automatic tool tracking and query augmentation.5-Minute Quickstart
Build your first memory-enabled agent
Self-Hosted Setup
Deploy Ryumem with Docker Compose
MCP Server
Add memory to Claude Desktop and coding agents
Core Concepts
Understand how Ryumem works
Open Source: Ryumem is open source under the AGPL-3.0 license. Self-host it on your own infrastructure or contribute on GitHub.
- API Server - FastAPI backend with REST API
- Dashboard - Next.js web UI for visualization
- MCP Server - Model Context Protocol server for AI agents
- Python SDK - Client library for Python applications
Key Features
Episode-Based Storage
Store conversations and facts as episodes that get automatically parsed into knowledge graphs
Smart Retrieval
Semantic + keyword + graph traversal search with temporal decay scoring
Auto-Deduplication
Embedding-based entity resolution prevents duplicate memories
Contradiction Handling
Automatically detects and resolves conflicting facts
Memory Pruning
Removes expired and obsolete information automatically
Web Dashboard
Modern Next.js UI with graph visualization and analytics