Batch Execute Retriever
Execute a retriever against multiple queries in a single request. The retriever is fetched and optimized once, then executed concurrently against each query with bounded parallelism.
Use case: IP safety / copyright clearance — scan 20 media files against face, logo, or audio retrievers in one call instead of 60 sequential SSE requests.
Limits: 1-50 queries per batch, 1-20 concurrency.
Returns results keyed by query index with per-query documents and errors.
Authorizations
Bearer authentication header of the form Bearer <token>, where <token> is your auth token.
Path Parameters
Retriever ID or name.
Query Parameters
Generate presigned URLs for S3-backed blobs and url-shaped fields. Also accepted as a body field — if either source is true, presigning is enabled.
Include vector embeddings in result documents. Also accepted as a body field — if either source is true, vectors are returned.
Body
Batch of queries to execute.
List of queries to execute (1-50). Each gets the same retriever pipeline.
1 - 50 elementsShared settings applied to every query. Supports 'limit' (max results per query, default 10) and 'max_chunks' (for content-mode preprocessing).
Max concurrent executions (1-20). Higher = faster but more resource usage.
1 <= x <= 20Generate presigned URLs for S3-backed blobs and url-shaped fields. Also accepted as a query param — if either source is true, presigning is enabled.
Include vector embeddings in result documents. Also accepted as a query param — if either source is true, vectors are returned.
Stream results via Server-Sent Events. Each query result is emitted as it completes, with keepalive pings every 15s to prevent proxy timeouts.
Response
Successful Response

