Everything you need to build intelligent, interactive knowledge graphs.
import { GraphStore, QueryEngine } from '@inferagraph/core/data';
const store = new GraphStore();
store.addNode('n1', { name: 'Alice', type: 'person' });
store.addNode('n2', { name: 'Bob', type: 'person' });
store.addEdge('e1', 'n1', 'n2', { type: 'knows' });
const query = new QueryEngine(store);
const neighbors = query.getNeighbors('n1', 1);
// → ['n2']Install InferaGraph and render your first interactive graph in minutes.
Work with nodes, edges, and domain-agnostic attributes using GraphStore.
Connect client-side: static datasets, REST APIs, GraphQL, or any custom backend.
Server-side backends: Gremlin, Cosmos DB, SQL, file systems, and Azure Log Analytics.
Keyword and semantic search with attribute filtering and graph traversal.
Pass an llm prop and unlock chat, semantic search, and tool-driven graph manipulation.
Stream conversational answers with useInferaGraphChat. Tool calls update the graph internally.
Three storage tiers behind one contract. Hash-stamped provenance, never re-embed unchanged content.
OpenAI, Anthropic+Voyage, Azure Foundry, mock, and a contract for building your own.
Pluggable response + embedding cache. lruCache built-in, Redis external, provider-bound invalidation.
Run the AIEngine in-process or behind an HTTP/SSE endpoint. Same Chat API either way.
Direct AIEngine.queryStream and provider stream() AsyncIterables — raw tokens without the chat layer.
Customize layouts, modes, themes, and node/tooltip rendering for stunning 3D graphs.
Click to inspect, hover to expand. Host-owned detail UI plus library-owned neighbor expansion.
Serialization, temporal navigation, clustering, multi-select, keyboard nav, and export.