This app is a demo of F5-TTS model that lets you clone any voice in your browser. No servers, no uploads, just pure client-side execution. You can feed a short audio of someone talking, and it will synthesize new speech in their voice saying whatever text you want. There’s also a podcast mode that lets you create multi-speaker conversation easily.
Once the models are downloaded, you are free to disconnect and work offline since everything runs locally and your audio never leaves your computer.