Skip to main content
Some Language Models support generating audio, processing audio inputs, or both. The following are examples for how to use the capabilities of these types of models in your application. For a list of models capable of audio input or output, please refer to the multimodal models on the Models page. On that page, you will can also lookup parameters which are model specific.

​
Transcribe audio into text

The following is an example of how you can use the audio input features to transcribe a sound file into text. 
import base64
import os

from opperai import Opper

opper = Opper(
    http_bearer=os.getenv("OPPER_API_KEY"),
)

def transcribe_audio(path: str):
    media_input = f"data:audio/mp3;base64,{base64.b64encode(open(path, 'rb').read()).decode('utf-8')}"
    response = opper.call(
        name="async_transcribe_audio",
        instructions="given an audio file, return the transcription of the audio",
        input={
            "_opper_media_input": media_input,  # this field name is required by opper
        },
        model="gcp/gemini-2.5-pro",
    )
    return response.message

if __name__ == "__main__":
    audio_path = "sample.mp3"
    transcription = transcribe_audio(audio_path)
    print(transcription)