AI Image
Ollama for Mac gains support for local text
9to5Mac • • general
Key Points:
- Ollama, a popular open-source platform for running AI models locally, has introduced text-to-speech generation support on its Mac app, with Windows and Linux versions coming soon.
- The new feature uses Phonon, a family of text-to-speech models by Gradium, enabling real-time speech generation in five languages and multiple voice accents.
- Ollama leverages Apple’s MLX framework on Apple Silicon Macs, achieving 35% faster speech generation compared to Gradium’s implementation while maintaining high accuracy.
- Users can download and run these speech models locally without internet access, ensuring privacy and efficiency, with support available in Ollama version 0.40.11 or later.
- The platform remains highly popular on GitHub, with nearly 185,000 stars, and allows users to access AI models from repositories like Hugging Face for various applications.
