Project JSON interface structure: TRANSCRIPTION jobs

Object detection job template

"TRANSCRIPTION_JOB": {
      "content": {
        "input": "<input type>"
      },
      "instruction": "<instruction>",
      "mlTask": "TRANSCRIPTION",
      "required": 1,
      "isChild": false,
      "autofill": false,
      "isNew": true
    }

For general information on the project JSON interface structure, refer to Generic project JSON interface template.

For information on how to use transcription in real-life labeling projects, refer to Transcription jobs.

Job-specific settings

Parameter	Value	Description
`mlTask`	`"TRANSCRIPTION"`	N/A
`input`	`"<input type>"`	Input type. One of: - `textField` for generic input - `number` for numeric input - `date` for date (in the YYYY-MM-DD format)
`autofill`	`true` or `false`	In Image, PDF and Text projects, you can define the behavior of a nested transcription task to either use or ignore the contents of its parent task. When `autofill` is set to `true`, the transcription will automatically be filled in with either of: `json_metadata` contents whose coordinates match the coordinates of the parent bounding box the contents of the parent NER annotation

Parameter

Value

Description

mlTask

"TRANSCRIPTION"

N/A

input

"<input type>"

Input type. One of:
- textField for generic input
- number for numeric input
- date for date (in the YYYY-MM-DD format)

autofill

true or false

In Image, PDF and Text projects, you can define the behavior of a nested transcription task to either use or ignore the contents of its parent task. When autofill is set to true, the transcription will automatically be filled in with either of:

json_metadata contents whose coordinates match the coordinates of the parent bounding box
the contents of the parent NER annotation

📘
The structure of the transcription type job does not contain any categories.

For information on how the transcription jobs look in the exported jsonResponse, refer to The structure of jsonResponse for exported transcription jobs.