Local Semantic File Search on macOS Using EmbeddingGemma 2

Developers have released a free, open-source search tool for macOS that allows users to find files based on their semantic description rather than just the filename.

The system utilizes Google's local EmbeddingGemma 2 model, ensuring high performance and complete data privacy without requiring registration or subscriptions.

The search engine can interpret user queries to locate various types of content:

• Photos (e.g., searching for 'dog on the beach' returns the corresponding image);
• Audio materials (finding specific moments in podcasts by context);
• PDF documents (locating relevant sections in contracts);
• Code (searching for functions based on their functionality description).

The DigUp utility runs completely autonomously on the user's device, eliminating dependence on cloud services.

You can download the latest version of the tool from the project page on GitHub: https://github.com/ARahim3/DigUp/releases/latest