Speech-to-text

openai/whisper-tiny-en

Model ID: openai/whisper-tiny-en
Type: Speech-to-text

Use this model

Transcribe audio in a web app

Install the browser SDK:

npm install @wfloat/wfloat-web

Load the model and pass a browser File, Blob, ArrayBuffer, or decoded audio buffer to transcribe:

import { loadSttModel } from "@wfloat/wfloat-web";

const stt = await loadSttModel("openai/whisper-tiny-en", {
  language: "en",
  onProgress(event) {
    console.log(event.status);
  },
});

const file = document.querySelector<HTMLInputElement>("#audio-file")?.files?.[0];
if (!file) throw new Error("Choose an audio file first.");

const result = await stt.transcribe({ audio: file });
console.log(result.text);

For microphone input, call await stt.startMicrophone(), then pass the value returned by await stt.stopMicrophone() to stt.transcribe(...).

Transcribe audio in a React Native app

Install the React Native SDK:

npm install @wfloat/react-native-wfloat

Run pod install from the app’s ios directory, and configure microphone permission for the target platform. Then record and transcribe locally:

import { loadSttModel } from '@wfloat/react-native-wfloat';

const stt = await loadSttModel('openai/whisper-tiny-en', {
  language: 'en',
});

await stt.startMicrophone({ sampleRate: 16000 });

// Call this when the user stops recording.
const recording = await stt.stopMicrophone();
const result = await stt.transcribe(recording);

console.log(result.text);

You can also call transcribe({ audio, sampleRate }) with mono PCM samples that your application already owns.

Transcribe audio in Python

Install the Python SDK:

pip install wfloat

Load the model and transcribe a mono 16-bit PCM WAV file:

import wfloat

stt = wfloat.load_stt_model(
    "openai/whisper-tiny-en",
    language="en",
)

result = stt.transcribe(audio="recording.wav")
print(result.text)

For in-memory audio, pass a sequence of floating-point mono samples and provide its sample_rate.