Output

OutputResponse and related structures returned from document processing

OutputResponse

The OutputResponse object contains the processed results of a document analysis task.

chunks · Chunk[] · required

Collection of document chunks, where each chunk contains one or more segments. See Chunk below.

file_name · string

The name of the file.

page_count · integer

The number of pages in the file.

pdf_url · string

The presigned URL of the PDF file.

extracted_json · object · deprecated

DEPRECATED: The extracted JSON from the document.

Chunk

A Chunk represents a logical grouping of segments from the document. Chunks are created based on the target_length configuration.

chunk_id · string · required

The unique identifier for the chunk.

chunk_length · integer · required

The total number of tokens in the chunk. Calculated by the configured tokenizer.

segments · Segment[] · required

Collection of document segments that form this chunk.

When target_chunk_length > 0, contains the maximum number of segments that fit within that length (segments remain intact). Otherwise, contains exactly one segment.

See Segment below.

embed · string

Suggested text to be embedded for the chunk. This text is generated by combining the embed content from each segment according to the configured embed sources (HTML, Markdown, LLM, or Content).

Can be configured using embed_sources in the SegmentProcessing configuration.

Segment

A Segment represents a logical element within a document page (e.g., title, paragraph, table, image).

segment_id · string · required

Unique identifier for the segment.

segment_type · SegmentType · required

The type of the segment. See Segment Types for all possible values.

bbox · BoundingBox · required

Bounding box coordinates for the segment.

Show BoundingBox properties

left · number · required

The left coordinate of the bounding box.

top · number · required

The top coordinate of the bounding box.

width · number · required

The width of the bounding box.

height · number · required

The height of the bounding box.

page_number · integer · required

Page number of the segment (1-indexed).

page_width · number · required

Width of the page containing the segment.

page_height · number · required

Height of the page containing the segment.

content · string · required

Content of the segment, will be either HTML or Markdown, depending on the format chosen in segment processing configuration.

html · string · required

HTML representation of the segment.

markdown · string · required

Markdown representation of the segment.

text · string · required

Text content of the segment. Calculated from the OCR results.

llm · string

LLM-generated representation of the segment. Only present if LLM processing is configured for this segment type.

image · string

Presigned URL to the cropped image of the segment. Only present if cropping is enabled for this segment type.

confidence · number

Confidence score of the layout analysis model for this segment (0.0 to 1.0).

ocr · OCRResult[]

OCR results for the segment.

Show OCRResult properties

text · string · required

The recognized text of the OCR result.

bbox · BoundingBox · required

Bounding box for this OCR result.

confidence · number

The confidence score of the recognized text (0.0 to 1.0).

Example Response

{
  "chunks": [
    {
      "chunk_id": "550e8400-e29b-41d4-a716-446655440000",
      "chunk_length": 256,
      "segments": [
        {
          "segment_id": "660e8400-e29b-41d4-a716-446655440001",
          "segment_type": "Title",
          "bbox": {
            "left": 72.0,
            "top": 100.0,
            "width": 450.0,
            "height": 36.0
          },
          "page_number": 1,
          "page_width": 612.0,
          "page_height": 792.0,
          "content": "# Document Title",
          "html": "<h1>Document Title</h1>",
          "markdown": "# Document Title",
          "text": "Document Title",
          "confidence": 0.95
        }
      ],
      "embed": "# Document Title"
    }
  ],
  "file_name": "example.pdf",
  "page_count": 10,
  "pdf_url": "https://s3.amazonaws.com/..."
}