local speech noise reduction
The future speech noise tool is designed for voice recordings with fans, traffic-like sound, office noise, and similar backgrounds. It will load its specialized noise-reduction module only when this task is chosen and should encourage a cleaned short preview before export.
Your selected file stays on this device. No upload, account, or server-side media job.
Browser-local workspace
Choose one local audio file—or a local video on the extraction route—to preview browser-readable facts and set one focused local plan. The source is not uploaded, modified, or sent to a fallback service.
No upload, account, or server-side media job.
The proposed path uses a narrowly loaded RNNoise-based processor, with Light, Recommended, and Strong controls intended to adjust suppression rather than simply lower overall volume. It is aimed at speech and is not a general music-restoration or dereverberation system.
Noise reduction is the only first-release Audio task planned to load a specialized WASM module. The design keeps that asset separate so opening a cutter, converter, or volume tool does not download speech-cleanup code.
Private browser handling
Aftermath reads a file you choose only in this browser tab. The shared workspace currently provides local file selection, playback, and task-specific planning; audio processing and export are not connected in this release.
How it works · Privacy & ad choices · Terms of use · Contact
