Audio processing
Audio editing, analysis, vocal isolation, speaker comparison and transcription in the browser toolbox run on your device. We do not upload your recordings, transcripts, subtitle edits or speaker embeddings to our servers. Microphone access requires your action and browser permission.
Models, speech voices and preferences
AI tools download model files and runtime code when first started. System text-to-speech uses your browser or operating system; a selected online voice may send text to that voice provider. Local and online voices are labeled. Language preferences and supported model caches are stored on this device and can be removed through browser storage controls.
Website usage analytics
We use FounderOmni Web Analytics to understand visits and improve the website. It records page paths, referrers, campaign tags, language, browser/device information and interaction events. It uses first-party local and session storage for visitor and session identifiers. The tracker respects Do Not Track and Global Privacy Control. Website analytics is separate from audio processing: the integration does not read audio file contents, transcripts or subtitle text.
Other network requests
Hosting infrastructure may process standard access logs. Google Fonts receives standard connection information when fonts load. Model providers receive requests for public model files. These network requests do not include audio uploaded into the local processing tools.
Protection limits and product status
VocalShield is experimental and does not promise to prevent every cloning or identification attack. Desktop protection is planned for audio routed through Audiomni; other apps may access your original microphone. Windows and macOS apps and the browser extension are in development. Free browser tools marked ready are available now.
Audiomni · A product by ArtyShield.ai
See models and open-source credits