Memento
Creates persistent memory for AI conversations using a SQLite knowledge graph that stores entities, observations, and relationships with semantic search capabilities.
9645 views4Local (stdio)
What it does
- Store entities, observations, and relationships in a knowledge graph
- Perform semantic vector search using BGE-M3 embeddings
- Retrieve contextually relevant information across conversations
- Search using both full-text and vector similarity
- Switch between SQLite and PostgreSQL backends
- Score relevance using temporal and contextual factors
Best for
AI assistants that need long-term memoryChatbots requiring context from previous conversationsKnowledge management systems with semantic searchApplications needing persistent entity relationships
Offline embedding model (no API calls)SQLite and PostgreSQL support1024-dimensional vector search
