Invoke the document retrieval pipeline (embedding + hybrid vector search + Reranking) and return matched document chunks.
Reranking: applied by default (skip_rerank=false). When active, each result includes a scoring.certainty field with the reranker confidence score. Set skip_rerank=true to return raw retrieval order without certainty.
Scoping: use workspace_id and/or tag_id to narrow retrieval to specific workspaces or tags, or use file_id to target specific files. file_id cannot be combined with workspace_id or tag_id (400). A 403 is returned if any provided filter resolves to no authorized resources. When no filters are provided, retrieval runs across all authorized documents.
Modes:
text (default): hybrid text searchvision: image-based searchBearer authentication header of the form Bearer <token>, where <token> is your auth token.
Natural-language search query.
Retrieval pipeline: "text" (hybrid search on DocumentChunk) or "vision" (image-based on VisionChunk).
text - textvision - visiontext, vision Number of chunks retrieved from the vector store before reranking. Range: 1–100.
1 <= x <= 100Number of chunks returned after filtering. Range: 1–50. Must be ≤ top_k.
1 <= x <= 50Scope retrieval to specific workspace IDs (authorized only).
Scope retrieval to specific file IDs (authorized only).
Scope retrieval to documents with any of these tag IDs (company-scoped).
Skip reranking. Useful to isolate retrieval quality.
Include base64-encoded page image in each result.
Chunks retrieved successfully. Empty array if no documents match.
Retrieved chunks with context