Transcriptions

Transcribe audio

Transcribes an upload minted by POST /api/v1/transcriptions/uploads and returns the text. Call it within 5 minutes of minting the upload. Credits are charged by file size. The uploaded audio is deleted when the call returns, whether it succeeds or fails, so each upload can be transcribed once.

Limit: this call is synchronous. The request times out after about 60 seconds, while the AWS speech backend can take up to about 5 minutes on long audio, so a long recording can fail with a gateway timeout (504) instead of returning text. A timed-out upload may not be deleted right away (it expires on its own), so do not retry with the same file_key: mint a new upload, or send shorter audio.

post/api/v1/transcriptions

Request

  • Base URL: https://your-deployment.example.com
  • URL: https://your-deployment.example.com/api/v1/transcriptions
  • Auth: one of:
    • HTTP bearer
    • API key in header x-api-key

Request body

file_keystring required

The file_key from POST /api/v1/transcriptions/uploads.

Example request

{
  "file_key": "transcribe-uploads/<userId>/<uuid>.mp3"
}

Response

The transcript.

textstring required

Example response

{
  "text": "Thanks for calling. How can I help?"
}

Changes

No recorded changes to this operation across all 1 revision of this API.