Product
Knowledge graph
The people, projects, and decisions in your work, connected - so asking about one brings back everything attached to it.
The knowledge graph runs in Neo4j with space-scoped Cypher so queries never bleed across tenants. Note, entity, and graph services keep captures addressable as linked objects instead of isolated files.
A Cypher query guard blocks prompt-injection patterns that could leak cross-space data. pgvector embeddings and hybrid buckets complement graph traversals for similarity and keyword recall.
Memory-service search and MCP assemble bundle vector hits with graph neighbors into one context package agents and the web UI can trust.
Outcomes
- Space-scoped Neo4j entities and relationships with guarded Cypher
- 1024-dim pgvector index paired with hybrid search buckets
- Graph plus vector context bundles for UI search and MCP assemble
Try it free
Start with knowledge graph on your next space-no credit card required on the free tier.
Related
Explore next
Product
Consolidate
Everything new gets read and linked to what you already knew - so when something contradicts an earlier decision, you hear about it.
Learn more →Platform
MCP & agent platform
Give your AI tools the same memory you have. They read what you know and save what they learn, inside the limits you set.
Learn more →Product
Capture & ingestion
Save anything from anywhere - type it, say it, snap it, or clip it. It all lands in one inbox, and the same thing twice never piles up.
Learn more →