VibeVoice

VibeVoice is an open-source voice AI covering both recognition and synthesis, developed in the open on GitHub. Voice models are customizable rather than fixed, and the project is built to sit alongside systems you already run instead of asking you to move onto a platform. Development happens publicly, so the code and the discussion around it are both visible, and outside contributions are part of how it moves. Developers, startups, and organizations building voice interaction are the ones who pick it up. It is a component you embed in your own product, not a finished service.