Returned type from polling a transcription

Returned transcript, present iff status == 'success'

The date and time when the transcription will no longer be available.
This date and time will be of a format like 'Jan 1, 1970 00:00:00 GMT'.
Present iff status == 'success'.


The accuracy score of the transcribed slice.
The accuracy score ranges from 0 to 1 and a higher score indicates a relatively higher confidence that the transcription is correct.


The duration of the slice in milliseconds

The start time of the slice in milliseconds

The accuracy score of the transcribed token.
The accuracy score ranges from 0 to 1 and a higher score indicates a relatively higher confidence that the transcription is correct.


Indicates whether the token has a definitive timestamp

The duration of the token in milliseconds

The start time of the token in milliseconds

A fully specified transcript with timing information

The accuracy score ranges from 0 to 1 and a higher score indicates a relatively higher confidence that the transcription is correct

Scribe's Batch API allows users to upload audio files for asynchronous processing.
Once a file has been uploaded, and unique, randomly generated ID is returned to the user.
Users can then use that ID to poll the Scribe Batch API to get results.

Accessing these APIs requires passing an authentication header in the standard format of
"Authorization: Bearer <YOUR_ACCESS_TOKEN>".

For individual users accessing the API, see the [authentication guide](../../authentication).


Scribe Batch API

This method takes in the unique_id returned when starting a transcription,
and deletes all stored data associated with that transcription job.


unique_id

Authentication header in the format of "Authorization: Bearer <YOUR_ACCESS_TOKEN>"


Authorization

All stored data associate with the transcription job is no longer available.
This could be due to a successful deletion or the transccription job never existed.
A 200 response does not contain any content body.


An invalid request was made to the server.
This could occur if the unique_id was not added as a path parameter or the unique_id is invalid.
A 400 response will contain a body describing the error.


The request was not authorized.
This could be due to credentials that are missing, expired, or invalid.
A 403 response will contain a body describing the error.


The transcription job for the request_id is still in progress.
This could occur if the transcription is still going on and that an attempt
at a later time will result in a 200 response.
A 409 response will contain a body describing the error.


Get the results of a transcription job.
This method takes in the unique ID returned from starting a transcription,
and returns either a JSON (default) or WebVTT response.

If the `Accept` header is missing or of type `application/json` then this call
will return a tuple of status, error, and transcript.
The status will be one of "success", "error", or "pending".
The transcript field will be present if and only if the status is "success".
The error field will be present if and only if the status is "error".
A status of "pending" indicates that the transcription job is still in progress,
in which case neither the transcript nor error will be present.

If the `Accept` header is of type `text/vtt` then this call will return
the transcription as VTT text tracks.


An invalid request was made to the server.
This could occur if the unique_id was not added as a path parameter.


The transcript for the request_id cannot be found.
It is possible that the transcription is still going on and that an attempt
at a later time will result in a 200 response.


Indicates that the request was rate limited.
This can occur if very frequent polling exists limits on the number of
allowed requests to the API.
A 429 response will contain a body describing why the upload was rate limited


An error occurred while processing the audio / video uploaded.
The response text will have additional details.


Starts a transcription job for a single audio or video file.

The `Content-Type` header in the request must be set to the type of audio or
video being uploaded. The `Content-Type` can be set to any one of the following
mime types that we support:
- `application/mp3`
- `audio/aac`
- `audio/m4a`
- `audio/mp3`
- `audio/mpeg`
- `audio/mpeg3`
- `audio/wav`
- `audio/wave`
- `audio/x-m4a`
- `audio/x-mpeg-3`
- `audio/x-wav`
- `video/mp4`

MP4 video, as a container, can hold any supported audio type.
MP3 and M4A audio support a wide variety of compression and audio codec options.

Files must be less than or equal to 1 gigabyte in size, and the upload must
take less than 30 minutes.

The body of the request can contain a single binary file OR multipart form data.  When
multipart form data is submitted there are a few additional options and requirements:
- The `Content-Type` header of the request must be `multipart/form-data`
- There must be file field with the name `file` which contains the audio / video data,
  the mime type for the data and an optional filename.
- There can be a form field with the name `TranscriptionOptions` which can contain
  transcription options like hotwords.


Indicates that the audio file was successfully uploaded.
This method returns a randomly generated unique ID that the client can
use to query the resulting transcript.


An invalid request was made to the server.
This could be because of an invalid content type, or an invalid MP3 file.
A 400 response will contain a body describing the error.


Indicates the upload was rate limited.
This can occur if audio is uploaded too quickly, or if a user's audio
quota is exceeded.
A 429 response will contain a body describing why the upload was rate limited


Batch API specification