Ollama is adding a new feature for Mac users that lets them generate speech from text locally, without a subscription. The update uses “Phonon,” a family of small text-to-speech models from Gradium, aiming for more natural-sounding voices while keeping processing on the device.

Both outlets describe the feature as part of Ollama’s broader focus on running AI models locally. Digital Trends emphasizes the availability and “free” nature of the local text-to-speech option on macOS, while 9to5Mac frames the announcement around expanding Ollama for Mac’s capabilities and notes that similar support is expected for other platforms soon, including Windows and Linux. Together, the reports position the release as an incremental upgrade to the Mac app rather than a change to Ollama’s overall approach.

No additional performance benchmarks, pricing tiers, or technical specifications are included in the provided summaries.