Response format

Learn about the response format for our parsing endpoint.

The parsing response format is optimized for flexibility with retrieval augmented generation.

See the entire response format in the API Reference.

Attribute breakdown

result · object

Show child attributes

result.type · string

Either “full” or “url” — defines how the chunks are delivered (inline or via URL).

result.chunks · array

Present only when type is “full” — an array of chunk objects.

chunks[].content · string

The content of the chunk extracted from the document.

chunks[].embed · string

Chunk content optimized for embedding and retrieval. For example, it differs from content when figure/table summaries are enabled.

chunks[].enriched · string | null

The enriched content of the chunk extracted from the document.

chunks[].blocks · array

Array of block-level structures—such as Text, Table, or Figure blocks.

blocks[].type · string (enum)

Block type indicating the content category: e.g., “Text”, “Table”, “Figure”.

blocks[].bbox · object

Contains bounding box information for the block, using normalized or spreadsheet-based coordinates.

blocks[].bbox.left · number

Normalized horizontal start coordinate (0–1 range), or column index in spreadsheets.

blocks[].bbox.top · number

Normalized vertical start coordinate (0–1 range), or row index in spreadsheets.

blocks[].bbox.width · number

Normalized width relative to page size (0–1 range), or column width count in spreadsheets.

blocks[].bbox.height · number

Normalized height relative to page size (0–1 range), or row height count in spreadsheets.

blocks[].bbox.page · integer

The parsed page number (1-indexed)—either page index or sheet index for spreadsheets.

blocks[].bbox.original_page · integer

The original page number from the source document, useful when filtering or slicing page ranges.

blocks[].content · string

The actual content of the block—text, table HTML, or figure caption/data.

blocks[].image_url · string|null

Presigned URL (if enabled) to download the block’s figure or table image; may expire (~24 h).

blocks[].confidence · string|null

Either a “low” or “high” confidence score for each block.

blocks[].logprobs_confidence · number|null

Numeric confidence score based on logprobs and OCR confidence.

result.url · string

Present only when type is “url” — endpoint to fetch the chunk JSON remotely.

Code snippet example

The Parse output can have two types: either full or url. The full type contains all the content and block information. Reducto swaps to using type=url when the request response size is close to exceeding the maximum allowed HTTP response size (6 MB for our current infra). Set force_url_result to true to consistently return URLs.

Full response example

{
  "job_id": "abcdef123456",
  "duration": 5,
  "pdf_url": null,
  "studio_link": "https://studio.reducto.ai/job/xyz",
  "usage": {
    "num_pages": 1,
    "credits": 0.5
  },
  "result": {
    "type": "full",
    "chunks": [
      {
        "content": "Detailed content for full LLM usage.",
        "embed": "Summarized content optimized for embeddings.",
        "enriched": "Additional semantic enrichment details.",
        "enrichment_success": true,
        "blocks": [
          {
            "type": "Header",
            "bbox": {
              // All bbox values normalized to [0,1] range
              "left": 0.1, // Distance from left edge
              "top": 0.05, // Distance from top edge
              "width": 0.8, // Width as % of page width
              "height": 0.1, // Height as % of page height
              "page": 1, // Current page number (1-indexed)
              "original_page": 10 // Original doc page number
            },
            "content": "Text content",
            "image_url": null, // presigned url to download figure/table image
            "confidence": "High",
            "logprobs_confidence": 0.92
          } ...
        ] ...
      } ...
    ]
  }
}