Speech-to-Text

Transcribe audio into text. STT runs as an asynchronous job: you enqueue the work, then poll the job until it finishes.

POST /v1/stt/jobs API key

Send either a multipart/form-data file or a JSON body with an audioUrl.

Fields

`file` * file	Audio file, up to 200 MB. (Or provide audioUrl in a JSON body.)
`audioUrl` uri	A URL to fetch the audio from, instead of uploading.
`language` string	Language hint, e.g. "lo". Auto-detected when omitted.
`projectId` uuid	Attach the job to a project.

Enqueue a job

curl -X POST https://api-ocr.vanguardinitiative.com/v1/stt/jobs \
  -H "authorization: Bearer $API_KEY" \
  -F file=@meeting.mp3 -F language=lo

const form = new FormData();
form.append('file', audioFile);
form.append('language', 'lo');
const res = await fetch('https://api-ocr.vanguardinitiative.com/v1/stt/jobs', {
  method: 'POST',
  headers: { authorization: 'Bearer ' + API_KEY },
  body: form,
});
const { jobId } = await res.json();

202 Job accepted

{
  "status": "success",
  "jobId": "b3f1c2a4-...",
  "state": "pending"
}

Poll for the result

Poll GET /v1/jobs/{id} until status is success or error.

curl https://api-ocr.vanguardinitiative.com/v1/jobs/$JOB_ID \
  -H "authorization: Bearer $API_KEY"

200 Finished job

{
  "status": "success",
  "job": {
    "id": "b3f1c2a4-...",
    "serviceId": "stt",
    "status": "success",
    "result": { "text": "ສະບາຍດີ ...", "language": "lo" }
  }
}