| Key features | - Automatic extraction and modeling from documents and system exports
- Source attribution on every fact with supporting text and confidence values
- EU-hosted or on-premises deployment
- Multi-hop query support for complex relationships
- RDF-based ontology with OWL 2 RL profile
- API and MCP access for AI agents
- Continuous graph updates and duplicate resolution
| - Browser-based WebAssembly engine with no server required
- Support for Cypher and GQL (ISO 39075) query languages
- Multiple deployment options: Docker, Linux, Windows, and embedded SDKs
- Studio IDE with query editor, graph view, schema browser, and algorithm tools
- Bolt protocol support (versions 5.1 to 6.0)
- HTTP API and gRPC for programmatic access
- Graph algorithms and APOC library support
- Embedded SDKs for .NET, Python, Java, Node.js, and Go
- Sample databases: Movies, Pokec social network, OurAirports, OpenFlights
|
|---|
| What makes it different | - Sourced, auditable facts for AI agents and people to verify before use
- Handles multi-hop questions and audit trails better than RAG alone
- Automatic infrastructure for knowledge graph building without manual pipeline setup
- Persistent knowledge graph with standard RDF/OWL exports for interoperability
| - Open source with MIT or Apache 2.0 license - deploy on your own infrastructure
- Runs in browser as WebAssembly with no installation or API key needed
- Same query engine and graph algorithms as the server version
- Production-tested infrastructure used by The Intelligence Company
- Drop-in replacement for Neo4j with compatible APIs
|
|---|
yggy
Traverse