Ollama can now read text aloud. It starts with Phonon
Ollama is adding support for voice models that turn written text into speech.
Ollama says that on computers with Apple Silicon chips, it generates speech more than a hundred times faster than real time. This is a vendor claim, not an independent measurement. Ollama blog The model comes from Gradium. Text-to-speech is available in five languages and with dozens of voices. Audio plays progressively as it is generated.