BLblink
AI

Transcribe audio to text

Transcribe audio to text using fal.ai Whisper. Provide either a public URL (audio_url) or base64-encoded bytes (audio_data). Billed per second of audio under ai:transcribe.

post/api/v1/ai/transcribe

Request

  • Base URL
    https://core.blink.new
  • URL
    https://core.blink.new/api/v1/ai/transcribe
  • Auth
    • HTTP bearer
    • or HTTP bearer

Request body

audio_urlstring uri

Public HTTP/HTTPS URL of the audio file to transcribe

audio_datastring

Base64-encoded audio bytes (for local files that cannot be publicly hosted). Mutually exclusive with audio_url — only one is required.

audio_mimestring

MIME type hint for audio_data (e.g. audio/mp3, audio/wav)

modelstring

Transcription model to use (defaults to fal-ai/whisper)

languagestring

ISO 639-1 language code hint (auto-detected when omitted)

promptstring

Optional context prompt to improve transcription accuracy

Example request

{
  "audio_url": "https://example.com/recording.mp3",
  "audio_mime": "audio/wav",
  "model": "fal-ai/whisper",
  "language": "en",
  "prompt": "Technical discussion about JavaScript frameworks"
}

Response

Audio transcribed successfully

textstring

Full transcribed text

languagestring

Detected or specified language code

durationnumber

Audio duration in seconds

Example response

{
  "text": "Hello, this is a test of the Blink transcription API.",
  "language": "en",
  "duration": 12.5
}

Changes

No recorded changes to this operation across all 1 revision of this API.