Speech to text

Convert speech to text with api-cohere-transcribe. Your API key must have access to this model.

Transcribe an audio file

Set the transcription model to api-cohere-transcribe.

Download speech.wav to your working directory, or use your own recording.

import os

from openai import OpenAI

client = OpenAI(
    api_key=os.environ["TRITONAI_API_KEY"],
    base_url="https://tritonai-api.ucsd.edu/v1",
)

with open("speech.wav", "rb") as audio_file:
    transcript = client.audio.transcriptions.create(
        model="api-cohere-transcribe",
        file=audio_file,
    )

print(transcript.text)

Read the response

The response contains a text field. The Python client exposes it as transcript.text.

This response came from a test WAV containing "The library opens at nine in the morning."

{
  "text": "The library opens at nine in the morning.",
  "usage": {
    "type": "duration",
    "seconds": 3.0
  }
}

Prepare the file

Use a non-empty recording. The example was tested with a mono, 16-bit PCM WAV at 16 kHz.

Keep audio within your approved data use. Do not send P4 information.

See the transcription reference for request fields and the response schema.