Voci

Voci is a local text-to-speech reader for Windows 10 and 11 that speaks in voices cloned from short audio recordings. It's for people who want to listen to documents or pasted text in a chosen voice without sending the content to a cloud service. After the initial downloads, it works offline and doesn't require an account or API key.

Coqui XTTS-v2 handles voice cloning and speech generation on your computer. You can save several named voices from WAV or MP3 recordings and use the same voice across supported languages, including English, Spanish, Arabic, Mandarin Chinese and Japanese. An NVIDIA GPU with CUDA support speeds up generation; CPU-only use works but is slower.

For long passages, playback starts while the app continues generating the remaining audio, rather than waiting for the whole document. You can pause, seek through generated audio, change speed or volume during playback, and cancel generation. Read-along highlighting follows the spoken words, though its timing is approximate. Text cleanup removes URLs, file paths, bracketed annotations and code fences that can make logs or chat transcripts awkward to hear.

The app uses a desktop window and stays available in the Windows system tray. It remembers playback and language settings between sessions. Voci's source code is free and MIT licensed; the speech model has a separate Coqui Public Model License that restricts it to non-commercial use.