downloadWhisperModel()v4.0.528

Downloads a Whisper model from remotion.media into the Transformers.js cache without initializing WebGPU. Call loadWhisperModel() or transcribe() afterwards.

The browser uses its Cache API; Node.js uses the Transformers.js filesystem cache. A cache must be enabled.

download-model.ts
import {downloadWhisperModel} from '@remotion/whisper-webgpu';

const {alreadyDownloaded} = await downloadWhisperModel({
  model: 'small.en',
  onProgress: ({progress}) => console.log(progress),
});

console.log(alreadyDownloaded);

Options

model

One of the names returned by getAvailableModels().

onProgress?

Called as files download. Default: no callback. The callback receives:

file

The file being downloaded, or null for aggregate progress.

progress

Aggregate progress between 0 and 1.

loadedBytes

Bytes downloaded in this call.

totalBytes

Estimated total model size in bytes.

Return value

alreadyDownloaded

true if all files were already cached before this call.

Compatibility

Browsers			Servers			Environments
Chrome	Firefox	Safari	Node.js	Bun	Serverless Functions	Client-side rendering	Server-side rendering	Player	Studio

Options​

model​

onProgress?​

file​

progress​

loadedBytes​

totalBytes​

Return value​

alreadyDownloaded​

Compatibility​

See also​