pngquant VS jpegoptim

image-shrinker is a simple, easy to use open source tool for shrinking images. Under the hood it uses pngquant, mozjpg, SVGO, and gifsicle. You can also install these tools individually if you need to compress some images. I often use pngquantafter exporting PNGs for web projects from Figma or similar tools. I literally run it like this:

Searching more I found https://pngquant.org/ which I could add to my bulk workflow to make most png's approach the jpeg size.

But this did prompt me to do some searching, and I see https://pngquant.org/ which seems to achieve jpeg like size reduction while maintaining the file as a png. One difference they note is that this method will typically preserve sharp edges better than jpeg (which is probably a strong plus for my type of use case)

If use PNG you can use pngquant 2 for instance. More modern image formats may also compress your images further.

I think it's likely the size is in the images - could you possibly call something like jpegoptim on the images as part of the build phase?

Maybe invoke jpegoptim as a subprocess, since it does lossless optimization of JPEG compression and I don't know of any crates that do that for JPEG.