HN Showcase
Back to browse
VocalSplit – Browser-based vocal remover running 100% locally

VocalSplit – Browser-based vocal remover running 100% locally

by jacoka·Mar 11, 2026·1 point·1 comment
Visit ProjectView on HN

AI Analysis

●●SolidWizardrySlick

WebAssembly ML inference beats server-based tools like vocalremover.org on privacy.

Strengths
  • •SIMD-optimized WebAssembly runs neural networks entirely client-side
  • •Parallel Web Workers utilize multi-core CPUs for faster processing
  • •No upload delays or file size limits since processing is local
Weaknesses
  • •Browser ML inference slower than native Demucs or Spleeter installations
  • •Vocal separation category already saturated with free and paid alternatives
Category
AI/ML
Target Audience

Musicians, content creators, podcasters needing vocal isolation

Similar To

vocalremover.org · Spleeter · Demucs

Post Description

I built VocalSplit because I wanted a way to separate vocals and music without uploading files to a remote server. Most existing tools are either paid SaaS or require a complex local Python environment (like Spleeter or Demucs).

This tool runs entirely in the browser. No data leaves your machine, and there are no processing limits or queues. It uses [említsd meg a technológiát, pl. WebAssembly vagy ONNX Runtime Web] to handle the inference locally.

Thanks for checking it out!

Similar Projects

AI/ML●●Solid

Building a Music Creation Platform Powered by Lyria 3

Text or an image -> a studio-style song with auto-generated lyrics and even a SynthID authenticity watermark; that's the product pitch and it's compelling in one line. The site bundles sensible add-ons (AI video maker, vocal remover, lyrics tool) that nudge it from a novelty toward a usable toolkit, but the space is crowded and the integration is still 'coming soon,' so the real test will be fidelity of Lyria 3 outputs and licensing/rights details.

SlickBold BetCrowd Pleaser
garmiter
103mo ago