Gepard 1.0 brings bilingual TTS with voice cloning
An open, autoregressive text-to-speech model targets English and Spanish under a permissive Apache 2.0 license.
A team publishing under the handle nineninesix has released Gepard 1.0, an autoregressive text-to-speech model that supports voice cloning across English and Spanish. The model arrives on Hugging Face under the permissive Apache 2.0 license, which allows both commercial and research use.
Gepard uses an autoregressive approach to generation, producing speech token by token rather than in a single pass. That design has become common in recent open TTS work, trading some efficiency for expressive, natural-sounding output and the flexibility to adapt to a target voice from a reference sample.
Why it matters
Open speech synthesis is a fast-moving corner of the AI ecosystem, and bilingual coverage remains relatively scarce compared to English-only releases. A model that handles both English and Spanish with cloning support lowers the barrier for developers building voice assistants, dubbing tools, and accessibility features across two of the world's most widely spoken languages.
A few practical notes for anyone evaluating the release:
- Languages: English and Spanish
- Capability: voice cloning from a reference sample
- License: Apache 2.0, permissive for commercial use
The repository does not publish detailed benchmark figures or parameter counts, so teams will want to test Gepard against their own audio quality and latency requirements before deploying. As an initial 1.0 release, it establishes a baseline the project can build on.
Sources
- Visit
nineninesix/gepard-1.0
Hugging Face
More in Text → Speech
Audio8 debuts a 0.6B multilingual zero-shot TTS preview
The compact text-to-speech model promises voice cloning across languages from a footprint small enough to run without heavy hardware.
KRAFTON releases A.X-K2 Raon speech MoE model
The game maker's new open model blends text-to-speech and speech recognition in a single 21B mixture-of-experts system with just 3B active parameters.
NVIDIA's Audex Unifies Audio Understanding and Speech
A new 30B mixture-of-experts model from NVIDIA handles both listening and speaking within a single audio-text architecture.
0 comments
No comments yet. Be the first to weigh in.