OpenVox setup and local voice workflows
A developer-led Windows tutorial covering OpenVox installation, voice cloning, EPUB audiobooks, model selection and a local MCP connection.
Player not loading? Watch on YouTube
OpenVox's developer demonstrates the speech app on Windows and describes availability for Mac and iPad. Setup includes a Windows app warning and a separate local voice runtime. The developer says processing works offline after model downloads.
The walkthrough covers model downloads, language selection, emotion tags and explicit pauses. Text settings handle numbers and abbreviations; audio settings include silence removal and loudness normalization. Batch generation supports an Excel template, while conversations assign voices to named speakers.
Voice cloning includes selecting a sample and checking Whisper's reference transcription. A separate Qwen3 TTS model creates voices from descriptions. The developer then generates speech with a designed voice using Pocket TTS and OmniVoice for comparison. He notes weaker accents and pronunciation in some less represented languages.
Chatterbox coverage includes Turbo and Multilingual options, cloning comparisons and a voice conversion demonstration. For audiobooks, the developer imports an EPUB, edits its chapters and starts generation with Kokoro.
The final demonstrations cover Select & Read shortcuts and connecting OpenVox's local MCP server to Codex. The tutorial closes with Starter and Pro limits. These demonstrations describe the app at recording time; model recommendations and quality judgments are the developer's.