Voice-to-Text#

Audio to text — for customer voice messages, call recordings, and any analysis of spoken content.

POST/v2/audio/transcribe

Request#

Field	Type	Meaning
`audio_base64`	string	Audio file as Base64 (MP3 or WAV)
`filename`	string	Optional, default `audio.wav`

bash

curl -s -X POST -H "Content-Type: application/json" -H "X-Tenant: bbe" \
  --data-binary @request.json \
  https://api.callfluence.staging.bestbrandseverywhere.com/v2/audio/transcribe

Response#

json

{
  "tenant": "bbe",
  "text": "Geschafft: Ihre Anfrage ist eingegangen. Sie erhalten in Kürze Ihren Rückruf unter der angegebenen Nummer.",
  "engine": { "engine": "cartesia", "mode": "together-fallback", "usableNow": true }
}

The engine self-report names the path actually used: Cartesia Ink directly (mode: "direct") or Whisper-v3 via the Together fallback.

Interplay with the transcription switch#

This endpoint is always available; whether ongoing calls are transcribed automatically is controlled by the tenant switch transcriptionEnabled in the audio settings.