DigUp includes an optional code search capability powered by EmbeddingGemma 2. Code search stays turned off until you activate it during setup or in Settings → Code. DigUp builds a separate index specifically for your repositories, so code entries never mix with standard desktop files. For general search controls, see the How to Use DigUp guide.
How to Query Code
To search your projects, prefix your query with code: in the search bar. You can describe what the code does or type a specific identifier name:
code: retry a failed call with backofffinds the function, whatever it is called.code: refreshAccessTokengoes to the line that defines it, not the places that call it.code: plot how many examples each class hasfinds the matching notebook cell.
Results open directly at the matching line in your editor. DigUp supports Zed, VS Code, Cursor, Sublime Text, JetBrains IDEs, Xcode on a Mac, and Notepad++ on Windows. You can choose your editor in Settings.
What Gets Indexed
DigUp reads source files, notebooks without their outputs, and a repository’s own documentation files.
- Skips files ignored by git.
- Skips vendored and generated code.
- Skips data files, keys, and passwords.
While you edit files, the index catches up one minute after things go quiet. On a Mac, reading code takes about 20 seconds for a small app, and about 20 minutes and 130 MB of index for all of llama.cpp. Settings displays how long reading each repository takes and how much room its index needs.