Growing voices for every language
VoiceGarden is cultivating a future where every language has a voice. We're creating high-quality, accessible text-to-speech voices for under-served languages, ensuring no community is left unheard.
Our Mission
Universal Access
Create text-to-speech solutions for all languages that are free for end users and "just work". We believe everyone deserves access to quality speech synthesis.
Open Development Pipeline
Build an open pipeline for developing new TTS voices using CommonVoice for data collection and Montreal Forced Aligner for phonemisation. Our comprehensive playbook helps communities create voices for their languages.
Immediate Impact
Deliver 6 new high-quality voices: 3 for underrepresented languages and 3 new variants for existing languages. This initiative demonstrates our commitment to linguistic diversity and inclusion.
VoiceGarden-SAPI
VoiceGarden-SAPI will make modern TTS voices available as standard Windows SAPI voices, so AAC, screen reader, and other Windows applications can use them without custom integration.
1,300+
offline languages through Sherpa-ONNX models
Azure
online neural TTS voices when cloud speech is available
SAPI
Windows system voice compatibility for existing apps
The code is working now and the next milestone is code signing, so Windows users can install it with confidence. Development happens in the open on GitHub.
What it enables
- Run Sherpa-ONNX voices offline on Windows for broad language coverage.
- Use Azure TTS online for high-quality cloud voices where available.
- Expose both routes through the Windows SAPI voice system.
- Bring more voices into AAC tools, screen readers, and legacy Windows speech apps.
Explore Available TTS Voices
Search and filter the current TTS voice catalog by language, provider, platform, and availability. The data is collected via py3-tts-wrapper, harmonized in the TTS-Dataset project, and served by the public TTS voice catalog API.
Accessibility First
Creating voices that work across all platforms and devices, ensuring everyone has access to quality speech synthesis.
Sustainable Voice Creation
Building open tools, documentation, and components that empower communities to create and maintain their own high-quality voices for generations to come.
Cultural Sensitivity
Working closely with communities to ensure voices are developed respectfully and authentically.
Join the Voice Garden Community
Be part of our mission to bring text-to-speech capabilities to every language. Sign up to receive updates about our progress and opportunities to contribute.