File STT

POSTapi.denthub.ai/sttasync job

Upload one or more audio files and have them transcribed as a background job. The request returns a job id straight away; the transcript is ready when you poll the retrieve endpoint below — or when your callbackUrl is notified.

Send up to 10 files in a single request by repeating the file part. Each file is transcribed and tracked independently.

File constraints

Maximum file size: 300MB
Maximum file duration: 120 minutes
Supported formats: WEBM, MPEG, WAV, AAC, FLAC, QUICKTIME, MP4

Bulk upload

Maximum files per request: 10
Maximum total request size: 3GB
Each file is validated independently against the per-file constraints above

Body parameters

Field	Type	Notes
client	string (uuid)	required — caller id
language	enum	required — ko \| en
file	file[]	required — 1–10 audio files (multipart)
speakerCount	integer	required — -1 auto-detect · 0 VAD (no diarization) · 1–4 fixed count
callbackUrl	string	optional — result webhook

GETapi.denthub.ai/stt/{sttId}poll result

When the job finishes, inspect data.files[]. Each file carries its own status and result.time_step_info[]; every segment includes contents, start and end times, language, and speakerId. For a single-file request,data.result mirrors the first file's result for convenience. See the poll response on the right for a representative per-file shape.

Example

Request

curl -X POST https://api.denthub.ai/stt \
  -H "X-License-Key: <license-key>" \
  -F client=123e4567-e89b-12d3-a456-426614174000 \
  -F language=ko \
  -F speakerCount=0 \
  -F file=@audio1.wav

Response · 200

{
  "data": { "id": "550e8400-e29b-41d4-a716-446655440000", "status": "PENDING" },
  "message": null
}

GET /stt/{sttId} · 200

{
  "data": {
    "licenseUsageId": "550e8400-e29b-41d4-a716-446655440000",
    "status": "COMPLETED",
    "requestType": "TRANSCRIBE_FILE",
    "files": [
      {
        "fileIndex": 0,
        "fileName": "audio1.wav",
        "status": "COMPLETED",
        "result": {
          "speaker_count": 2,
          "time_step_info": [
            {
              "contents": "Hello, this is speaker one.",
              "start_time": 0.0,
              "end_time": 1.86,
              "language": "en",
              "speakerId": "1"
            },
            {
              "contents": "And this is speaker two.",
              "start_time": 1.86,
              "end_time": 3.42,
              "language": "en",
              "speakerId": "2"
            }
          ]
        }
      }
    ],
    "resultLatency": 1666.0
  },
  "message": null
}

Errors & authall STT & LLM REST endpoints

Authenticate every REST call with the license key issued to your account, in the X-License-Key header. Errors are returned as an ErrorResponseEntity:

{
  "errorCode": "1003",
  "message": "Invalid License."
}

HTTP	errorCode	Meaning	When
400	1000	Bad Request	Request validation failed or malformed request body
401	1003	Unauthorized	Missing or invalid license key in X-License-Key header
404	1002	Not Found	Requested resource not found or access denied
500	5000	Server Error	Internal server error — retry later or contact support

Some endpoints run as background jobs and return a status you poll (or receive by webhook); others respond synchronously. See Asynchronous jobs for how polling and callbacks work.