LAVU
Local AI photo search.
The problem
Finding photos inside a large local collection is surprisingly difficult. You remember a face, a place, or the feeling of a photo — but rarely the filename or the folder it ended up in.
Your library should be easy to explore without having to upload it to someone else’s servers.
The solution
LAVU lets you search your local photo library using images, faces, and AI. It brings the search to your photos, keeping your library on your own computer.
What it does
Image search
Start with a photo and find visually similar images in your library. Follow the visual connection instead of guessing a filename.
Face search
Find photos of the same person across your collection, even when they’re scattered through different folders.
Local-first
Your photo library stays on your computer. Search is designed around local processing, without requiring a cloud photo upload.
Multiple AI models
Use models suited to different tasks: visual understanding for image search, and face representations for finding people.
A product in progress
LAVU is in active development. This page will grow with product screenshots, release notes, and lessons from building it.
Built with
- Electron
- PyTorch
- DINOv2
- ArcFace
- FastAPI