Article · AI translation
IT Home reported on October 27 that Ollama, an open-source platform for running AI models locally, announced on October 24 that its Mac app now officially supports text-to-speech (TTS) generation, with Windows and Linux versions to follow.
Technology outlet 9to5Mac considers this Ollama's first addition of speech generation, taking local AI a step further from being able to “read and write” to being able to “speak.”
The first supported speech model is the Phonon series developed by Gradium. The model is relatively small and has advanced technical specifications. Users can download it directly through Ollama and run it on an Apple Silicon Mac without relying on cloud services, making it better suited to situations where privacy and offline use matter.
On macOS, Ollama uses Apple's open-source machine learning framework MLX to take full advantage of Apple Silicon's unified memory architecture and improve operating efficiency.
Official tests show that, on a MacBook Pro equipped with an M5 Max chip, Phonon generates speech 35% faster than Gradium's own implementation while maintaining a word error rate comparable to that of the Hugging Face Transformers reference implementation.
The new feature also supports real-time speech generation. Users do not have to wait for the AI to finish producing text; they can hear the speech as the response is generated, making the interaction feel closer to a natural conversation.
Citing a blog post, IT Home reported that Phonon currently offers separate models for 5 languages—English, French, German, Spanish and Portuguese—as well as a multilingual model supporting all 5. Users can also choose voices with different accents, such as American or British English.
Ollama is one of the popular open-source AI projects on GitHub, with nearly 185,000 stars and about 18,600 forks. Users can download and run open AI models locally or access more powerful cloud models as needed.