About SoundScrub
SoundScrub is a desktop application that uses advanced AI to remove unwanted sounds from videos with precision and ease.
What is SoundScrub?
SoundScrub is a professional audio isolation tool designed for video creators, editors, podcasters, and content producers who need to control specific sounds in their video files. Unlike traditional audio editing software that requires manual selection and filtering, SoundScrub uses AI-powered source separation to intelligently isolate any sound you describe. Remove it completely, keep only it, or adjust the volume balance between the targeted sound and everything else.
The application extracts audio from your video file, sends it to the cloud for AI processing, then merges the processed audio back into your original video. You can remove background music, people speaking, ambient noise, or any custom sound you describe in natural language. Your video file stays on your machine throughout the process.
How It Works
1. Select Your Video
Select any video file from your computer. The app supports all common video formats and automatically extracts the audio for processing. Your video file stays on your machine.
2. Pick a Sound and Choose What to Do
Use preset options like speech, background music, or ambient noise - or describe any custom sound using natural language. Then choose your mode: remove the sound completely, keep only it, or adjust the volume of each layer independently with sliders.
3. AI Processing
The app extracts audio from your video and sends it to the cloud for processing. SoundScrub's AI analyzes the audio and separates the unwanted sounds. The processed audio is then returned to your app. Your original video never leaves your machine.
4. Get Your Video Back
Once processing is complete, the app downloads the processed audio and merges it back into your original video. In adjust mode, the volume mixing happens locally on your machine. Video and audio quality remains unchanged - only the sound balance is different.
Common Use Cases
Content Creators
Remove copyrighted background music, eliminate unwanted dialogue, or turn down background noise without losing it entirely using adjust mode.
Podcasters
Clean up ambient noise, remove background conversations, eliminate HVAC hum, or strip out unwanted music from interview recordings.
Video Editors
Isolate dialogue from location sound, remove unwanted ambient noise, or use adjust mode to dial down specific sounds while keeping them in the mix.
Musicians
Extract or remove specific instruments from recordings, isolate vocal tracks, or remove drum tracks to create custom backing tracks.
Educators
Remove background noise from lecture recordings, eliminate unwanted student chatter, or clean up classroom recording environments.
Archivists
Restore historical recordings by removing tape hiss, background interference, or other artifacts while preserving the primary audio.
Powered by Meta's SAM Audio
SoundScrub is built on top of SAM Audio (Segment Anything Model for Audio), a state-of-the-art audio source separation model developed by Meta Research. SAM Audio uses text-prompted audio separation - you describe a sound in natural language, and the model isolates or removes it from the audio mix.
While SAM Audio is a powerful research model, using it directly requires Python, GPU setup, and command-line expertise. SoundScrub makes SAM Audio accessible to everyone by wrapping it into a simple desktop app - no coding, no terminal, no environment setup. Just drag your video in, pick a sound to remove, and get your clean video back.
For developers looking to integrate SAM Audio into their own applications, check out the SAM Audio API.
Technical Capabilities
AI-Powered Source Separation: Powered by Meta's SAM Audio model, SoundScrub uses advanced machine learning trained on millions of audio samples to identify and isolate specific sound sources within complex audio mixtures.
Natural Language Processing: Describe any sound you want removed in plain English. SAM Audio understands context and can target specific audio elements based on your description.
Three Output Modes: Remove a sound completely, keep only it, or use adjust mode to control the volume of the targeted sound and everything else with independent sliders.
High-Quality Output: Original video quality is preserved. Only the audio track is modified, maintaining video resolution, frame rate, and encoding.
Cross-Platform Support: Available for macOS, Windows, and Linux. Audio processing happens in the cloud while your video stays on your machine.
Job History & Resume: Track all your processing jobs, resume interrupted downloads, and maintain a complete history of processed videos.
Pricing Model
SoundScrub uses a simple, transparent credit-based pricing system. You pay only for what you process:
$0.20 per 30 seconds
A 5-minute video costs $2.00. A 30-minute video costs $12.00. Credits never expire and can be used anytime.
No subscriptions, no hidden fees. Buy credits as you need them and use them at your own pace.
Privacy & Security
Your videos never leave your computer. Only the extracted audio is sent to our servers for processing, and it's deleted immediately after processing is complete. We don't store, analyze, or use your content for any purpose other than performing the requested audio removal.
All communication between the desktop app and our processing servers is encrypted. Your payment information is handled securely by Stripe, our payment processor - we never see or store your payment details.