Whisper Turbo is a fast, cross-platform Whisper implementation, designed to run entirely client-side in your browser/electron app.
With Whisper Turbo, you can add transcription to any app in minutes.
Check out Getting Started for more.
WebGPU is only officially supported on Chromium based browsers running on Windows & MacOS. For Linux support, check here.
Install whisper-turbo:
npm install whisper-turbo
const session = useRef<InferenceSession | null>(null);
const loadModel = async () => {
//The session manager handles constructing the inference session.
const manager = new SessionManager();
const loadResult = await manager.loadModel(
AvailableModels.WHISPER_TINY,
() => { console.log("loaded!") },
(progress: number) => { console.log("Loading: ", progress) }
);
if (loadResult.isErr) {
console.log("Failed to load!");
} else {
session.current = loadResult.value;
}
};
const runSession = async () => {
await session.current.transcribe(
your_uint8_array,
true/false,
(s) => {
console.log("Segment!")
}
);
};
Coming soon
- Are you a GPU wizard?
- Do you know what a HRTB is in Rust?
- Do you know what is going on here?
- Reach out: [email protected]