Response Format - TextIn Intelligent Document Parsing

Parse returns Markdown, structured document elements, and optional output fields based on the capabilities you enable in the request. This page describes the standard synchronous response format. If you download the final result from an asynchronous result_url, the file contains the parse result object directly, without the outer code, message, and data wrapper.

Response overview

{
  "code": 200,
  "message": "success",
  "data": {
    "schema_version": "1.3.0",
    "file_id": "doc_7f3a2b",
    "metadata": {},
    "markdown": "# Title\n\nContent...",
    "elements": [],
    "pages": [],
    "title_tree": [],
    "summary": {}
  }
}

Top-level fields

Field	Type	Description
`code`	integer	Application-level response code
`message`	string	Response message
`data`	object	Parse result
`x_request_id`	string	Request identifier

`data` fields

Field	Type	Description
`schema_version`	string	Response schema version
`file_id`	string	File identifier
`job_id`	string	Task identifier, when returned
`success_count`	integer	Number of pages processed successfully
`metadata`	object	File-level metadata
`markdown`	string	Document content in Markdown format
`elements`	array	Structured document elements
`pages`	array	Page-level information, returned when the `pages` capability is enabled
`title_tree`	array	Document outline, returned when the `title_tree` capability is enabled
`summary`	object	Processing summary

`metadata`

The metadata object contains file-level information. Common fields include:

Field	Type	Description
`filename`	string	Original file name
`filetype`	string	File type or extension
`page_count`	integer	Number of pages in the document
`data_source`	string	Source type, when returned

`elements`

The elements array contains structured document elements in reading order. Each element represents a structural unit in the document, such as a heading, paragraph, table, image, formula, header, or footer.

Basic structure

{
  "elements": [
    {
      "element_id": "el_001",
      "type": "Title",
      "text": "Annual Report 2024",
      "page_number": 1,
      "coordinates": [0.1, 0.12, 0.9, 0.12, 0.9, 0.16, 0.1, 0.16],
      "metadata": {}
    }
  ]
}

Base fields

Field	Type	Description
`element_id`	string	Unique identifier for the element
`type`	string	Element type
`sub_type`	string	Optional subtype
`text`	string	Recognized text
`page_number`	integer	Page number, starting at 1
`coordinates`	array	Normalized four-point bounding box
`metadata`	object	Element metadata
`objects`	array	Inline objects, returned when `include_inline_objects` is enabled
`table_structure`	object	Structured table data, returned for Table elements when `include_table_structure` is enabled
`char_details`	array	Character-level output, returned when `include_char_details` is enabled
`image_data`	object	Image output, returned for Image elements when `include_image_data` is enabled

Element types

Type	Description
`Title`	Heading
`NarrativeText`	Body paragraph
`ListItem`	List item
`Table`	Table
`TableCaption`	Table caption
`Image`	Image
`FigureCaption`	Figure caption
`Formula`	Math formula
`Chemical`	Chemical formula
`Header`	Page header
`Footer`	Page footer
`PageNumber`	Printed page number
`PageBreak`	Page break
`CodeSnippet`	Code block
`UncategorizedText`	Uncategorized text

`elements.metadata`

An element’s metadata object can include structural and relationship fields. Common fields include:

Field	Type	Description
`parent_id`	string	Parent element identifier
`children_ids`	array	Child element identifiers
`category_depth`	integer	Depth in the heading or section hierarchy
`ref_element_id`	string	Referenced element identifier
`is_continuation`	boolean	Whether this element continues from a previous element
`continuation_of`	string	Identifier of the earlier element this one continues
`has_inline_objects`	boolean	Whether inline objects are present
`inline_object_types`	array	Types of inline objects included in the element
`width`	number	Element width, when returned
`height`	number	Element height, when returned
`data_source`	string	Element source type, when returned

Coordinate system

Parse element coordinates are normalized to the page width and height. Coordinates use eight values:

[x1, y1, x2, y2, x3, y3, x4, y4]

The points are ordered as:

(x1,y1) -------- (x2,y2)
   |                |
   |  Element Area  |
   |                |
(x4,y4) -------- (x3,y3)

Each value is in the range 0 to 1. To convert normalized coordinates to pixel coordinates, multiply x values by the page width and y values by the page height.

page_width = 1576
page_height = 1683

x1_px = coordinates[0] * page_width
y1_px = coordinates[1] * page_height

`table_structure`

When include_table_structure is enabled, Table elements can include structured row, column, and cell data.

{
  "type": "Table",
  "text": "<table>...</table>",
  "table_structure": {
    "rows": 3,
    "cols": 2,
    "cells": [
      {
        "cell_id": "cell_1_1",
        "row": 1,
        "col": 1,
        "row_span": 1,
        "col_span": 1,
        "text": "Product"
      }
    ]
  }
}

Cell fields

Field	Type	Description
`cell_id`	string	Cell identifier
`row`	integer	Row index
`col`	integer	Column index
`row_span`	integer	Number of rows spanned
`col_span`	integer	Number of columns spanned
`text`	string	Cell text
`coordinates`	array	Cell coordinates, when returned
`content_type`	string	Cell content type, when returned
`image_datas`	array	Image content in the cell, when returned
`objects`	array	Inline objects in the cell, when returned
`char_details`	array	Character-level output for the cell, when returned

Use table_view: "html" when Markdown does not preserve the table structure well.

`image_data`

When include_image_data is enabled, Image elements can include extracted image data. Common fields include:

Field	Type	Description
`image_url`	string	URL for the extracted image
`mime_type`	string	MIME type
`base64`	string	Base64-encoded image data, when returned

`char_details`

When include_char_details is enabled, elements can include character-level details. Common fields include:

Field	Type	Description
`index`	integer	Character index in the element
`text`	string	Character text
`coordinates`	array	Character coordinates
`recognition.confidence`	number	Character confidence score
`recognition.candidates`	array	Candidate recognition results, when returned

`objects`

When include_inline_objects is enabled, elements can include inline objects such as formulas, checkboxes, handwriting, or inline images. Common fields include:

Field	Type	Description
`object_id`	string	Object identifier
`type`	string	Object type
`text_range`	object	Start and end positions in the parent text
`image_data`	object	Object image payload, when returned
`metadata.display_mode`	string	Display mode metadata, when returned

`title_tree`

The title_tree field is returned when the title_tree capability is enabled. Use it to build:

Table of contents
Section navigation
Structure-aware chunks

`pages`

The pages field is returned when the pages capability is enabled. Each item in pages can include:

Field	Type	Description
`page_number`	integer	Page number, starting at 1
`page_width`	integer	Page width
`page_height`	integer	Page height
`angle`	integer	Page rotation angle
`element_ids`	array	Element identifiers on the page
`page_image_url`	string	Page image URL, when returned
`dpi`	integer	Page DPI value, when available
`status`	string	Page processing status

`summary`

The summary object contains processing metrics. Common metrics include:

Field	Type	Description
`duration_ms`	integer	Total processing time in milliseconds

Parse-specific errors

Most authentication, file, and request validation errors are shared across Parse and Extract. See Errors for the full common error reference. The following error is specific to Parse:

Code	Description	Solution
`40427`	Unsupported DPI value	Use a supported DPI setting, or remove the DPI parameter if it is not required.

Parse Quickstart

Parse your first document.

Configuration

Customize parsing behavior.

Async Processing

Process large documents asynchronously.

Errors

Review common error codes.

​Response overview

​Top-level fields

​data fields

​metadata

​elements

​Basic structure

​Base fields

​Element types

​elements.metadata

​Coordinate system

​table_structure

​Cell fields

​image_data

​char_details

​objects

​title_tree

​pages

​summary

​Parse-specific errors

​Related resources