Hybrid Search Chunks

Hybrid search across chunks using BM25 + vector search.

Combines keyword matching (BM25) with semantic similarity (vector embeddings) using Reciprocal Rank Fusion to rank results. Searches across all documents the user has access to within their company.

Returns chunks with relevance scores, ordered by combined ranking.

get/api/v1/chunks/search

Request

The document declares no server URL.
Auth: none declared

Query parameters

querystring required

Search query text

document_idsinteger[] nullable

Filter by document IDs

matrix_idinteger nullable

Filter by matrix ID

entity_set_idinteger nullable

Filter by entity set ID

skipinteger

Number of results to skip

limitinteger

Maximum results to return

use_vectorboolean

Enable vector search (default: true)

Headers

authorizationstring nullable

x-api-keystring nullable

Response

Successful Response

querystring required

Search query text

total_countinteger required

Total number of results found

has_moreboolean required

Whether more results are available

Changes

No recorded changes to this operation across all 5 revisions of this API.