Voxlumo API v1

Health

GET /api/v1/health

Transcribe

Send multipart/form-data with an audio file. Use language=auto for automatic language detection, or pass ar, en or de for the current web-supported presets.

POST /api/v1/transcribe
Content-Type: multipart/form-data

audio=<file>
language=auto

Example response

{
  "transcript": "Plain transcript text…",
  "speakerTranscript": "[00:03–00:08] A: Hello…",
  "summary": "Business-ready summary…",
  "segments": [
    {
      "id": "seg_1",
      "speaker": "A",
      "start": 3.1,
      "end": 8.4,
      "startLabel": "00:03",
      "endLabel": "00:08",
      "text": "Hello…"
    }
  ],
  "speakers": ["A", "B"],
  "srt": "1\n00:00:03,100 --> 00:00:08,400…",
  "vtt": "WEBVTT…",
  "duration": 124.7,
  "capabilities": {
    "timestamps": true,
    "diarization": true,
    "subtitles": true
  }
}

Response fields

• transcript — plain transcript without speaker/timestamp decoration.
• speakerTranscript — readable transcript with speaker labels and time ranges.
• segments — structured diarized timeline with numeric start/end seconds.
• speakers — unique speaker labels detected in the recording.
• srt / vtt — subtitle-ready payloads generated from the diarized timeline.
• duration — detected duration in seconds when returned by the transcription engine.
• capabilities — explicit booleans indicating which structured outputs are present.

Current pilot limits

• The direct API endpoint accepts up to 25 MB per request.
• The Voxlumo web workspace can prepare larger recordings in the browser and process them in speech-quality sections; this large-file browser workflow is separate from the direct API contract.
• Supported direct formats: FLAC, MP3, M4A, WAV, MP4, MPEG/MPGA, OGG and WEBM.
• This public pilot endpoint does not advertise enterprise authentication, guaranteed SLA, batch queues or dedicated data-region controls.

Enterprise integration path

For a production integration, the next controls are API authentication, rate limits, asynchronous jobs for long audio, customer-specific retention, auditability and contractual security requirements. These should be implemented to the customer's actual requirements rather than claimed in advance.