Knowledge Graph
Connections, not just embeddings.
Every conversation builds a web of entities and relationships. When you ask a question, graph traversal finds connections that vector search misses.
Knowledge Graph
Connections, not just embeddings.
m3mory builds an entity graph from every conversation. People, places, projects, preferences. All connected. All searchable.
How It Works
Entity Extraction
People, places, organisations, projects, preferences, events. Automatically identified and stored as nodes in your knowledge graph.
Relationship Mapping
Typed edges connect entities. When your user mentions a colleague, a project, or a preference, the graph captures the relationship.
Graph Traversal
When you ask about a person, the graph finds everything connected to them: their projects, preferences, colleagues, and recent events.
Temporal Tracking
When facts change, the graph updates. Old information is archived, new information takes priority. Your AI always has the latest context.
Why graphs matter
Vector search finds memories with similar meaning. But it struggles with multi-hop queries: "What project is Alex working on, and what tech does it use?"
A knowledge graph handles this naturally. It traverses from Alex to their project, then from the project to its technology stack. Two hops, one query, precise results.
Combined with our multi-signal search pipeline, graph traversal gives m3mory a significant edge on relationship-heavy queries that pure embedding systems miss.