Mac version of Ollama adds offline text-to-speech, up to 35% faster than the original solution
Ollama announced on October 24 that its Mac app officially supports text-to-speech generation, with Windows and Linux versions to follow. The first supported speech model is the Phonon series developed by Gradium, which users can run offline on Apple Silicon Macs without cloud services. Official tests show that on a MacBook Pro with the M5 Max chip, Phonon, while achieving a word error rate comparable to the Hugging Face Transformers reference implementation, generates speech 35% faster than Gradium's own implementation.
Why read this
AI score 58
Ollama brings a local TTS model to Mac, and provides word error rate and speed comparisons with the reference implementation, which can be used to judge the usability of offline speech generation.
Topics
Body · AI translation
Mac version of Ollama adds offline AI speech generation: data never leaves the device, up to 35% faster than the original solution
2026-10-27 09:00 IT Home - Guyuan
IT Home, October 27 news: Ollama, the 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.
Tech media outlet 9to5Mac believes this is the first time Ollama has added speech generation capabilities, taking local AI from "able to read and write" a step further toward "able to speak."
The first supported speech model is the Phonon series developed by Gradium. The model is small in size with advanced technical specifications; users do not need to rely on cloud services and can download it directly through Ollama and run it on Apple Silicon Macs, making it better suited to privacy-conscious, offline use cases.
On macOS, Ollama uses Apple's open-source machine learning framework MLX, fully leveraging the unified memory architecture of Apple Silicon to improve operating efficiency.
Official tests show that on a MacBook Pro equipped with the M5 Max chip, while maintaining a word error rate comparable to the Hugging Face Transformers reference implementation, Phonon's speech generation speed is 35% faster than Gradium's own implementation.
Citing a blog post, IT Home explains that Phonon currently offers separate models for five languages—English, French, German, Spanish and Portuguese—along with a multilingual model supporting all five languages; users can also choose voices with different accents such as American English or British English.
Ollama is one of the most 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, and can also connect to more powerful cloud models as needed.