@fugood/node-whisper-wasm

Browser WASM implementation package for @fugood/whisper.node. Application code should import @fugood/whisper.node; this package is pulled in by the main package for browser builds.

The package exposes the same high-level context API as the native packages, but model and audio file paths are fetched as URLs and copied into the Emscripten filesystem before inference.

import { initWhisper } from '@fugood/whisper.node'

const whisper = await initWhisper({
  filePath: 'https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-base.en.bin',
  useGpu: false,
})

const { promise } = whisper.transcribeFile('https://raw.githubusercontent.com/ggml-org/whisper.cpp/master/samples/jfk.wav', {
  language: 'en',
})

console.log(await promise)
await whisper.release()

The package ships both single-thread and pthread WASM artifacts. On cross-origin isolated pages with SharedArrayBuffer, the loader uses the pthread artifact; otherwise it falls back to the single-thread artifact automatically. Whisper transcription defaults to up to 8 threads based on browser hardware concurrency when pthreads are available; pass maxThreads to override it. Browser WASM clamps maxThreads to the compiled pool limit of 8, or 1 in the single-thread fallback. Browser pages run model loading, transcription, benchmarks, and VAD in a dedicated module worker by default so the UI thread can keep rendering. Use the main whisper.node package entrypoint in browser code:

import { configureWasm, initWhisper } from '@fugood/whisper.node'

Use configureWasm({ worker: false }) only when you explicitly need the in-thread runtime, configureWasm({ threads: false }) to force the single-thread artifact, or pass workerPath, jsPath, and wasmPath when serving the package files from custom URLs. The older workerUrl and runtimeScriptUrl option names still work. Model downloads are cached in browser Cache Storage by default. Pass cacheModel: false to disable persistent caching, modelCacheName to isolate the cache namespace, or modelCacheKey when the fetch URL is a proxy or signed URL but should reuse the same cached model. Set useGpu: true only with a package built using GGML_WEBGPU=ON and a browser that exposes navigator.gpu. VAD currently falls back to CPU in the browser package because the Silero VAD graph hits unsupported WebGPU ops.

The default build emits wasm/whisper-node.js, wasm/whisper-node.wasm, wasm/whisper-node.threads.js, and wasm/whisper-node.threads.wasm. Use bash scripts/build-wasm.sh --single-file only when you want the WASM binary embedded into each generated JS file. Modern Emscripten embeds the pthread worker bootstrap in the main JS file, so a separate whisper-node.worker.js is not expected. The package module worker that keeps UI work off the main thread is worker.js. Use npm run build-wasm-docker to build through Docker; it selects a native arm64 Emscripten image on Apple Silicon hosts.