Ollama can now turn text into speech on your Mac, starting with Gradium's Phonon
Ollama, the popular app for running AI models on your own computer, now does text-to-speech, starting with Gradium's small open Phonon models. It runs on Apple Silicon Macs today, with Windows and Linux support announced as coming soon.
What it does
After installing Ollama 0.40.11 or later, ollama run phonon "text" speaks through your speakers, and adding > speech.wav saves a file; --voice picks one of Gradium's voices, and ollama show phonon lists them. Audio streams as it is generated, you can pipe a chat model's reply straight into Phonon, and an OpenAI-compatible /v1/audio/speech endpoint lets existing code switch by changing the base URL. Models cover English, French, German, Spanish and Portuguese, plus one multilingual model.
Results and limits
Ollama says that on a MacBook Pro with M5 Max it ran Phonon 35% faster than the next fastest engine, Gradium's own runtime, measured as the median of 10 runs generating 45 seconds of speech. This is Ollama's own benchmark. The library page lists 121 MB of loaded memory for the 31M English model and 141 MB for the 40M models, with 24 kHz mono output.
Worth knowing
Only the default 31M English model is Apache 2.0; the 40M English, French, German, Spanish, Portuguese and multilingual models are CC BY-NC 4.0, so not for commercial voiceovers. Phonon itself was released by Gradium on 21 October; Ollama's packages need Apple Silicon for now.
Try it yourself
ollama run phonon --voice Zoey "Welcome back to the channel." > voiceover.wav