Skip to main content

​
Create Transcription

Transcribes audio into text using the specified model. The request must be sent as multipart/form-data. 
POST https://api.runcrate.ai/v1/audio/transcriptions

​
Request Body (multipart/form-data)

ParameterTypeRequiredDescription
modelstringYesThe transcription model to use (e.g., openai/whisper-large-v3-turbo).
filefileYesThe audio file to transcribe. Supported formats include mp3, wav, m4a, webm, and mp4.

​
Example Request

curl https://api.runcrate.ai/v1/audio/transcriptions \
  -H "Authorization: Bearer rc_live_YOUR_API_KEY" \
  -F "model=openai/whisper-large-v3-turbo" \
  -F "file=@recording.mp3"

​
Response

{
  "text": "Welcome to Runcrate. The fastest way to run AI in the cloud.",
  "duration": 4.32,
  "language": "en"
}
FieldTypeDescription
textstringThe transcribed text.
durationnumberDuration of the audio file in seconds.
languagestringDetected language code (ISO 639-1).

​
Python Example

import requests

with open("recording.mp3", "rb") as audio_file:
    response = requests.post(
        "https://api.runcrate.ai/v1/audio/transcriptions",
        headers={
            "Authorization": "Bearer rc_live_YOUR_API_KEY",
        },
        files={
            "file": ("recording.mp3", audio_file, "audio/mpeg"),
        },
        data={
            "model": "openai/whisper-large-v3-turbo",
        },
    )

result = response.json()
print(result["text"])