AI audio cleaner for voice recordings that removes background noise, filler words, and silences in your browser
Free screen recorder that automatically polishes recordings with zooms, smooth cursors, captions and multi-track editing
Stage
Startup
Not given
Founded
2024
Not given
Based in
RO
Not given
Pricing model
Freemium
Open source
Pricing
Free Trial: 30 minutes of processed audio with all cleaning tools included. Starter: $9/month for 10 hours of processed audio per month. Unlimited: $20/month for unlimited processing on every tool.
Not given
Key features
Background noise removal with AI detection
Filler word removal in 20+ languages
Audio enhancement for studio sound
Silence and dead air removal
Mouth sounds and breath removal
Podcast transcription and summarization
Multitrack editing for multiple speakers
Timeline export of removed segments
MCP integration for AI agents
Support for multiple audio formats
Multi-track timeline with independent tracks for zooms, captions, shortcuts and annotations
Automatic zoom detection following clicks and typing
Smoothed cursor with click ripple effects and resizing
Keyboard shortcut visualization as keycaps on timeline
Live drawing and annotation tools during recording
Background blur, replacement and face-following stickers
Dynamic aspect ratio switching between wide, square and vertical
On-device caption generation with transcript-based editing
AI agent integration via Model Context Protocol
Motion blur on fast camera movements
What makes it different
No software installation required; everything runs in browser
Free tier available to try all cleaning tools without credit card
Processes most audio files in under 30 seconds
Removes filler words in 20+ languages
AI agents integration via MCP protocol
Everything runs locally on your computer with no accounts, uploads or telemetry
Built with native tools for each system for performance and integration
Open JSON project format works on both macOS and Windows
Completely free and open source with MIT license
Crash-safe recording that preserves work on power loss