Roadmap
Models
- XTTS - Production-ready multilingual TTS
- Kokoro - StyleTTS 2-based English TTS without voice cloning
- F5-TTS - Superb voice cloning and naturalness, but slower and less stable
- Parler TTS - Describe a voice with a text prompt
- StyleTTS 2 - Fast and efficient zero-shot voice cloning
- CosyVoice2 - Zero-shot voice cloning
- MetaVoice - 1.1B parameter zero-shot voice cloning model
- Fish Speech 1.5 - Zero-shot voice cloning
- OpenVoice V2 - Open source zero-shot voice cloning by MyShell
Features
- Simple Python API for easy integration
- Command-line interface for quick testing and batch processing
- REST API and web interface for remote access
- Model benchmarking tools
- Batch processing support
- Audio post-processing options
- Allow easier extensibility with a plugin system