Skip to main content
At Camb AI, we believe in democratizing access to powerful voice technology. We’ve released significant portions of our research as open-source models, allowing the community to build, experiment, and innovate.

​
MARS 6 Turbo

MARS6: A Small and Robust Hierarchical-Codec Text-to-Speech Model MARS 6 Turbo represents the latest evolution in our open-source speech synthesis technology. It delivers robust performance with a hierarchical codec architecture, optimized for efficiency without compromising on quality.

​
Resources

MARS 6 Turbo GitHub

Access the model weights, inference code, and technical documentation.

​
MARS 5

MARS 5 remains a powerful tool in our open-source suite, offering a transformer-based architecture for flexible speech generation.

​
Key Features

  • Flexible Prosody: Context-aware rhythm and intonation.
  • Voice Cloning: Zero-shot cloning with short reference audio.
  • Multilingual Support: Diverse language and accent capabilities.

MARS 5 GitHub Repository

Check out the MARS 5 code and models.

​
Community & Contribution

We welcome contributions! Whether it’s adding support for new languages, optimizing inference, or building new applications on top of MARS, join us on GitHub to shape the future of open-source speech AI.