Ollama for Mac gains support for local text-to-speech models

Ollama, the widely-used application for running AI models locally, has announced support for text-to-speech generation on its Mac version, with Windows and Linux support expected soon. The platform, which enables users to download and run open AI models on their Macs or access more powerful models via the cloud, is also available for Windows and Linux.

Ollama is open-source and has gained significant popularity on GitHub, boasting nearly 185,000 stars and 18,600 forks. It leverages Apple's MLX framework on macOS, enhancing the performance of supported AI models. Users can now download compatible open-source AI models from repositories like Hugging Face and run them directly on their Macs without needing an internet connection or sending data to external servers.

The introduction of Phonon, a family of text-to-speech models from Gradium, marks the first time Ollama supports speech generation. Phonon offers five languages (English, French, German, Spanish, and Portuguese) and a multilingual model, along with various voices and accents. The new speech models on macOS are 35% faster than Gradium's own implementation, according to tests on a MacBook Pro with M5 Max, while maintaining the same word error rate.

Users can now listen to AI-generated responses in real-time as they are generated. The new speech models are available for Ollama 0.40.11 or later on Apple Silicon Macs. Users can access Phonon through Ollama and find further information, audio samples, and instructions at a provided link.