Response Format - TextIn Intelligent Document Parsing

The API returns JSON. On success, the processed image is returned as a Base64-encoded JPG inside result.image_list, along with the original and cropped dimensions, the crop corner coordinates, and the detected orientation angle.

Response overview

The image value below is truncated for readability. The real response contains the full Base64-encoded JPG.

{
  "code": 200,
  "message": "success",
  "msg": "success",
  "version": "v2.0.8",
  "duration": 100,
  "x_request_id": "7596b8c9d2ddbc9924b66651e9efc174",
  "result": {
    "origin_width": 2000,
    "origin_height": 3000,
    "image_list": [
      {
        "cropped_width": 1500,
        "cropped_height": 1800,
        "image": "/9j/4AAQSkZJRgABAQAAAQABAAD/2wBD",
        "position": [0, 10, 500, 10, 500, 300, 0, 300],
        "angle": 90
      }
    ]
  }
}

Top-level fields

Field	Type	Description
`code`	integer	Application-level response code. The API always responds with HTTP 200; `200` here indicates success. See Errors for other codes.
`message`	string	Response message
`msg`	string	Response message (duplicate of `message`)
`version`	string	API version number, such as `v2.0.8`
`duration`	number	Processing time, in milliseconds (ms)
`x_request_id`	string	Unique request identifier, present in every response
`result`	object	Processing result

`result` fields

Field	Type	Description
`origin_width`	integer	Width of the original image
`origin_height`	integer	Height of the original image
`image_list`	array	Processed image results

`image_list` fields

Each item in image_list represents one processed image. image_list always contains a single item: for image input there is one result, and for PDF input only the first page is processed.

Field	Type	Description
`cropped_width`	integer	Width of the processed image
`cropped_height`	integer	Height of the processed image
`image`	string	Processed image as a JPG, Base64-encoded. Omitted when `only_position=1`.
`position`	array	The four corner points of the cropped region (8 values). See Crop coordinates.
`angle`	integer	Orientation angle. Always present. See Orientation angle.

Crop coordinates

position is an array of 8 values describing the four corner points of the cropped region in the original image:

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

The points are ordered as:

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

Index	Point	Corner
`0, 1`	(x1, y1)	Top-left
`2, 3`	(x2, y2)	Top-right
`4, 5`	(x3, y3)	Bottom-right
`6, 7`	(x4, y4)	Bottom-left

Coordinates are in pixels, relative to the original image.

Orientation angle

angle is always present in each image_list item. It reports the detected orientation of the input document, using the values below. Orientation is detected only when correct_direction=1; otherwise, or when detection fails, the value is -1.

Value	Orientation
`0`	Upright (▲)
`90`	Rotated right (▶)
`180`	Upside down (▼)
`270`	Rotated left (◀)
`-1`	Not detected or detection failed

Save the processed image

The image field is a Base64-encoded JPG. It is raw Base64 without a data:image/jpeg;base64, prefix, so add the prefix yourself if you embed it directly in an HTML <img> tag. Decode it to write the file to disk:

import base64

image_base64 = result["result"]["image_list"][0]["image"]

with open("output.jpg", "wb") as f:
    f.write(base64.b64decode(image_base64))

Quickstart

Process your first image.

Configuration

Control cropping, correction, and enhancement.

API Reference

Full request and response schema with an interactive playground.

Errors

Error codes and how to resolve them.

​Response overview

​Top-level fields

​result fields

​image_list fields

​Crop coordinates

​Orientation angle

​Save the processed image

​Related resources

Quickstart

Configuration

API Reference

Errors

Response overview

Top-level fields

`result` fields

`image_list` fields

Crop coordinates

Orientation angle

Save the processed image

Related resources