New Open-Source Tool Deletes 12GB of Apple Intelligence Data on macOS
Original Source: The Verge Apple
•
Read time: 2 min read
•Published: October 5, 2026
Share:
Source: The Verge Apple
Executive Summary
A new open-source command-line tool, RemoveMacAI, allows macOS users to easily delete up to 12GB of Apple Intelligence data. This tool addresses the removal of a dedicated toggle in macOS 27 that previously disabled Apple's AI features. Essentially, RemoveMacAI restores the functionality of the old toggle, helping users reclaim storage space and manage their AI features.
**New Open-Source Tool Deletes 12GB of Apple Intelligence Data on macOS**
Mac users looking to reclaim valuable storage space might find a new solution in an open-source command-line tool named `RemoveMacAI`. This utility promises to simplify the process of deleting up to 12GB of data associated with Apple Intelligence features on macOS.
The need for such a tool arises from a recent change in Apple's operating system. Previously, users could easily disable Apple Intelligence through a single toggle in Settings. However, this convenient option was removed in macOS 27. Consequently, the underlying AI models now persist on the user's disk, and their associated features are dispersed across various settings panes, making manual removal cumbersome.
As initially highlighted by *MacRumors*, the developer of `RemoveMacAI` explained on Reddit that the tool effectively replicates the functionality of the deprecated settings toggle. It systematically deletes the stored AI models and deactivates the related features. The developer noted the complexity of manual removal, stating, "You have to find about a dozen settings, and several of them are under Screen Time."
By offering a streamlined method to manage Apple Intelligence data, `RemoveMacAI` provides a practical solution for users who wish to optimize their Mac's storage and control the presence of these AI features on their devices.
Confirm and follow the full story at the original source:The Verge Apple
Found this interesting? Share it with your network: