neumann-graph

What problem does it solve?

Enable developers and analysts to store, model, and query directed property graph data alongside tables and vectors in a single Neumann instance, removing the need to synchronize separate graph and vector systems.

Core Features & Use Cases

• Graph modeling and CRUD: Create, read, update, and delete nodes and edges with labels, properties, and auto-generated IDs.
• Traversal and pathfinding: Neighbor queries, shortest and weighted paths, variable-length paths, and cross-engine neighbor ranking by vector similarity.
• Algorithms & analytics: Run PageRank, Betweenness, Closeness, Eigenvector, Louvain, and Label Propagation for graph insights; supports constraints, indexes, aggregation, and pattern matching.
• Cypher support and batching: Use a Cypher subset for complex queries and perform batch create/update/delete operations for scale.
• Use case example: Build a knowledge graph with entities and embeddings to power RAG, agent memory, or permissioned graph queries that respect relationships and semantic similarity.

Quick Start

Create a person node named Alice with role engineer and then run a NEIGHBORS query to find her outgoing connections.