
DigUp
Search photos, PDFs, recordings, videos, and code by what they contain, right on your Mac.
DigUp is a free, open-source search app for macOS that finds files by what's inside them, not just by filename. It uses Google DeepMind's EmbeddingGemma 2 model on your Mac to search pictures, PDFs, documents, recordings, videos, and optionally source code. Searches combine natural-language descriptions with exact text matches for names, numbers, and other specific terms.
Find a scene, passage, or moment
- Photos and screenshots: Search for what an image shows, or for text recognized inside screenshots.
- PDFs and documents: Find relevant passages, including in scanned PDFs, and open PDF results in Preview.
- Audio and video: Search for something said in a recording or shown in a video, then jump to the matching moment.
- Across languages: Search across languages supported by the model, although accuracy varies.
A customizable Shift–Command–Space shortcut opens the search panel from any app. Results can be previewed, revealed in Finder, or viewed in a larger window with a picture grid and filters.
Search code separately
Code search is optional and off by default. Enable it for selected project folders, then use the code: prefix to look for a function by what it does or by its name. Results can open at the relevant line in editors such as VS Code, Cursor, Zed, Xcode, and JetBrains IDEs. Code has a separate index and does not appear in ordinary file searches.
Local indexing and setup
DigUp indexes only folders you choose and does not upload indexed files or require an account. Setup downloads roughly 865 MB of model files. Initial indexing can take time for large folders; subsequent changes are indexed incrementally. Files stored only in iCloud are not downloaded for indexing.
The macOS app requires macOS 14 or later on Apple Silicon; Intel Macs are not supported. Install it from GitHub Releases or the developer's Homebrew tap.