Hybrid search
POST /data/v1/query/hybrid
Combines semantic similarity search with knowledge graph relationships. Extracts entities from the query using a dictionary-based extractor, then merges and re-scores vector + graph results with configurable weighting.
Authorizations
Request Body required
object
The search query text
How do I use Python SDK for embeddings?Weight for semantic search results in final scoring
Weight for knowledge graph results in final scoring
Filter results to specific sources
Entity names to boost in results ranking
Whether to extract entities from the query before searching
Game domain for entity extraction (e.g. ‘eve’, ‘valorant’)
Whether to use Neo4j graph traversal for entity expansion
Maximum graph traversal depth when use_graph is enabled
Responses
200
Successful response
object
Entity names extracted from the query
A single result from hybrid search with source attribution
object
Combined relevance score (0-1)
Source of this result
Number of results from vector search
Number of results from graph traversal
Number of results after merging and deduplication
400
Invalid request
object
object
401
Invalid or missing API key