Credits & licences
VDecoy is an original interface and story player, using openly licensed speech technology. Generated audio is synthetic.
Speech engine
Local TTS by Tristan McInnis, based on PocketTTS-RAVEN by Pantelis Kalogiros and PocketTTS.cpp by VolgaGerm. Runtime code is MIT licensed. This app uses source commit b20d545e173db2c01eb08a80bb9411b4dd6fed19, with modified loaders for same-origin model parts and integrity checks.
Voice model
Pocket TTS by Kyutai Labs, CC BY 4.0. The included model files are optimised ONNX derivatives. Record only your own voice here, and identify generated speech as synthetic. The app's WAV exports include an attribution and synthetic-audio comment.
Runtime components
The WebAssembly runtime includes ONNX Runtime from Microsoft and its third-party components. Attribution and complete licence notices are provided below. No external fonts, tracking libraries or cloud voice services are used by this app.