Audio

Audio Transcription

Upload an audio file as multipart form data. The gateway routes it to an enabled audio transcription model and returns the transcript with model and token usage details.

Endpoint

POST https://aiservice.prabisha.com/transcript

Multipart Request

Use file as a file field. Do not manually set the multipart Content-Type; let the HTTP client add the boundary. Audio files are limited to 15 MB.

curl -X POST https://aiservice.prabisha.com/transcript \
  -H "X-API-Key: YOUR_API_KEY" \
  -F "file=@./meeting.mp3;type=audio/mpeg" \
  -F "prompt=Transcribe the audio and preserve speaker labels"

Form Fields

FieldTypeDetails
fileFile, requiredAudio file. Supported MIME types include WAV, MP3/MPEG, AIFF, AAC, OGG, FLAC, M4A, L16, OPUS, ALAW, and MULAW.
promptstringOptional transcription instruction. Defaults to returning only the spoken words.
modelstringOptional enabled audio model ID. An enabled transcription-capable model is selected when omitted.

Response

{
  "success": true,
  "data": {
    "transcript": "The recognized words from the audio.",
    "model": "gemini-3.5-transcribe",
    "providerName": "GEMINI",
    "providerId": "provider-id",
    "providerModelId": "model-id",
    "usage": {
      "promptTokens": 100,
      "completionTokens": 20,
      "totalTokens": 120
    },
    "latency": 1200,
    "fallbackChain": ["GEMINI"]
  }
}