Mole developer Tw93 says the app’s recent updates came from testing software one installation at a time, with AI assistance, rather than relying solely on general cleanup rules. In an article published on November 13, 2026, Tw93 describes work spanning Mole versions 1.13 through 1.19, including AI-tool and macOS system-data cleanup. The Windows desktop version was released on October 14.
Tw93 reports testing 853 Mac apps to improve uninstall and leftover-file detection. An initial round of more than 300 apps led to 13 reusable rule categories, more than 90 specific cleanup paths and over 100 unit tests. It also identified more than 30 locations that Mole had treated as safe to clean but should leave alone. According to Tw93, an AI agent handled much of the installation, checking and removal process, while Tw93 intervened where needed.
For the AI cleanup feature, Tw93 says they tested more than 30 tools, including 13 desktop apps and 13 command-line tools. The resulting options distinguish between disposable data and potentially valuable work: users can choose retention periods of 15, 30, 90 or 120 days for old sessions, with a recovery copy made before deletion. Worktrees appear for cleanup only when Git can establish that they contain no uncommitted changes. For opencode.db, Mole compacts the database rather than deleting conversations.
The system-data feature takes a similarly cautious approach. Tw93 says Mole reports the size of items it can identify and handle, rather than presenting macOS’s broad “System Data” figure as a deletion target. Large items such as iPhone backups and virtual machines are shown with management links but cannot be selected for bulk cleanup.
Tw93 also reports testing 482 Windows apps for the desktop version, mostly in a Parallels virtual machine and partly on a remotely connected ThinkPad X1. These figures and the described results are the developer’s account of their testing, not an independent assessment of Mole’s cleanup performance.