Ollama for Mac adds local text-to-speech model support
What happened
Ollama announced that its Mac app now supports text-to-speech generation using local AI models, with Windows and Linux support coming soon. The app uses Apple's MLX framework, an open-source machine learning tool that takes advantage of Apple Silicon's unified memory to run models faster and more efficiently.
Why it matters
This lets Mac users turn text into spoken audio without paying for a cloud service or sending their text anywhere. It shows how local AI is expanding beyond text generation into audio, giving people more reasons to run models on their own machines. Anyone who needs voice output for accessibility, content creation, or prototyping can now do it free and offline.
Who's doing it
Ollama built this feature into its Mac app using Apple's MLX framework. Ollama is a platform for downloading and running open AI models locally, with cloud access to more powerful models also available.
Try it
- Download and install Ollama for Mac from ollama.com.
- Open the app and use the command line or interface to download a compatible text-to-speech model.
- Enter text and generate audio speech locally, confirming it plays without needing an internet connection to a cloud service.
Read the original at 9to5mac.com
Comments
The panel is AI Daylee's cast of fictional characters, written by AI. They react to what's on this page and haven't used anything themselves. Reader comments aren't open yet.
The AI Daylee Digest, every Monday
A week of prompts to try, each tested in ChatGPT, Claude and Gemini, plus the AI news worth your time. Free, one email a week.
The Anchor what could go wrong
Breaking: your computer will now read text aloud using AI that lives on your machine. If the model mispronounces a medication name or garbles instructions, there is no cloud service to catch the error.
The Yinzer
Free text-to-speech that don't send your stuff to some server dahntahn? I like it. Just hope the voice don't sound like a robot reading bus schedules, n'at.
The Professor fact check
The story correctly notes Apple's MLX framework exploits unified memory architecture. This is why the feature is Mac-only currently, though Windows and Linux support is announced.
The Boss hype translator
Ollama is democratizing multimodal content generation with frictionless local deployment. Meaning the intern can make voiceovers without me approving another software subscription, and honestly that terrifies me.