What is Lap?
Lap is a desktop photo manager for macOS, Windows and Linux that works on the folders you already have instead of importing them into a closed library. It indexes large collections, more than 100,000 files according to the README, in a local SQLite database.
Its AI runs on your machine through ONNX Runtime: CLIP lets you search by describing a photo or by picking a similar one, with optional search in more than 50 languages, and InsightFace detects and clusters faces so you can name people and browse their photos. It also reads more than 60 photo, RAW and video formats, finds duplicates and keeps thumbnails browsable when an external drive is unplugged.
Lap is built with Tauri, Rust and Vue, licensed GPL-3.0, and has about 4,400 GitHub stars.
Is Lap worth it?
Lap does the two things people miss most when they leave Google Photos or iCloud: find a picture by describing it, and see every photo of one person. It does both offline, on folders you already have.
The approach is sensible. Rather than copying your library into its own database, Lap indexes folders in place and keeps metadata and thumbnails in a local SQLite file, so you can still browse an external drive's library when the drive is unplugged. CLIP handles text and similarity search, InsightFace handles faces, and both run through ONNX Runtime on your own hardware; nothing is uploaded.
It covers the dull jobs too. More than 60 photo, RAW and video formats are supported, Smart Albums save rule-based views, and duplicate clean-up shows how much space you would reclaim before deleting.
The limits are those of a young project. Version 0.3.3 shipped on 12 October 2026, there are about 90 open issues, and the indexing speed and accuracy on a 100,000-photo library will depend on your machine. CLIP search finds scenes and objects well but is weaker on fine detail, and Lap does not yet write captions; two community forks add that with a local vision model. Compared with Immich, Lap has no phone backup or server, which is a limit if you want one shared family library and a strength if you want one simple desktop app.
Where it earns its place
A family archivist with 80,000 photos on an external drive
Find 'kids at the beach with a red bucket' without scrolling for an hour
Point Lap at the photo folders, let it index once, then search by description and filter by named faces. It is free; the cost is a long first indexing run.
A photographer who shoots RAW
Clear duplicates and find shots by subject across years of shoots
Index the archive, run the duplicate finder to see reclaimable space, and use similarity search from a reference frame. Files stay where they are.
Before you commit
- Setup
- Download the installer for macOS, Windows or Linux from Releases, add your photo folders and wait for the first index.
- Integrations
- Works on local and external folders; community forks add local captioning through an OpenAI-compatible server such as Ollama or LM Studio.
- Data and privacy
- Everything stays on your machine: a local SQLite database, models run through ONNX Runtime, no upload. GPL-3.0.
- Cost at scale
- Free; the cost is disk space for thumbnails and the time to index large libraries.
- Limits
- About 90 open issues; v0.3 release line; no mobile app or sync; search quality depends on the CLIP model.
What’s good
- Search photos by description or similarity, fully offline
- Local face clustering with named people
- Works on existing folders; no import into a closed library
- Over 60 photo, RAW and video formats, plus duplicate clean-up
What’s not
- Young project at v0.3
- No phone backup or multi-user server
- No built-in captions; forks add them
- Indexing speed depends on your hardware
Best for
People with large photo collections on their own disks who want AI search and face grouping without uploading anything.
Lap: common questions
- Is Lap worth it?
- Google Photos-style search by description and by face for pictures on your own disks, entirely offline, in a young project that is already useful. We rate it 4.1 out of 5.
- How much does Lap cost?
- Free and open source under GPL-3.0.
- What are the downsides of Lap?
- Young project at v0.3; no phone backup or multi-user server; no built-in captions; forks add them; indexing speed depends on your hardware.
- Who is Lap best for?
- People with large photo collections on their own disks who want AI search and face grouping without uploading anything.
- How much does Lap cost for a team or in production?
- Free; the cost is disk space for thumbnails and the time to index large libraries.
- Is Lap safe for company data?
- Everything stays on your machine: a local SQLite database, models run through ONNX Runtime, no upload. GPL-3.0.
Answers checked Nov 15, 2026.