Apple is reportedly developing a new Mac app called “Harbor” that helps users run local AI models on one or more Macs. The app would set up a private AI inference server, allowing large language models to run with data kept on-device rather than sent to external services.
Across outlets, the emphasis is on making local inference easier and more scalable. AppleInsider describes “Harbor” as previously known as “Fabric,” framing it as a more user-friendly way to create a local inference server and run open-weight models on the Mac itself. 9to5Mac focuses on the discovery of the app’s existence through leaker evidence, highlighting that Harbor is intended to make running local AI models across multiple Macs straightforward.
