Lighterpix
About Lighterpix
Why Lighterpix exists, how it works, what it collects (nothing) and what it is built on.
The idea
Image compression tools usually work by uploading your file to someone else’s computer. That is a strange thing to do with a passport scan, a medical letter or an unreleased product shot. Lighterpix does the same job with code that runs on the machine the image is already on.
How it works
The page loads a set of encoders compiled to WebAssembly — MozJPEG, OxiPNG, libwebp, libavif, libjxl — and runs them inside a Web Worker. The file you drop is read with the File API and never leaves the tab. With the network panel open you can watch: after the first load there are no uploads at all.
What it collects
No accounts, no cookies for advertising, no analytics that identify you, no third-party requests. The service worker caches the app so it keeps working offline.
What it is built on
Lighterpix is a derivative of the Squoosh project, which Google released under the Apache License 2.0. The original copyright notices stay in the source files, attribution is recorded in the NOTICE file, and the encoders keep their own licences — see THIRD-PARTY.md in the repository for the full list.