Free Mac App Searches Every Photo and Video Frame Locally

A developer released a free Mac app called SCM on GitHub, the code-sharing site. It searches by meaning across every photo and video frame stored on disk. All processing happens locally. There is no account, no cloud, and no uploading files anywhere. The search runs on CLIP and SigLIP, computer vision models that match images to their meaning instead of just matching file names. The default model weighs about 435 megabytes. It processes one frame in 50 to 570 milliseconds, depending on which version you pick.

For searching a personal photo and video library, a model like this is enough. It is a simple, repetitive task, and small local models handle it well. Writing code is a different story; that needs much bigger models. The app is free and does not require a subscription for cloud search, since all processing stays on the user's own disk. To install it, you need to run three separate commands one after another: brew tap allenv0/scm, then brew trust allenv0/scm, then brew install --cask allenv0/scm/scm.