Serving brotli compressed files

Site :

I’m trying to speed my website by using brotli static compression. It works fine in Safari (content-type is br in response headers) but there’s no mention of it in chrome and firefox. I was wondering if it was working fine because using compression didn’t improve my pagespeed lighthouse score. Any tips?

Hi @NicolasK

I see (via visiting it) no issues with the load time of the site. Using also shows good speed (this is mobile) shows slightly higher for mobile performance, and even higher for desktop.

I don’t see this is Safari. For Nicolas-Sursock-small.jpg I see Content-Type: image/jpeg. I do see br in the Accept-Encoding request header Accept-Encoding: gzip, deflate, br (as is also present in Chrome)

This is the only image I can see that comes from Netlify. All other images come from Unsplash which in the case of Safari are Content-Type: image/webp and Chrome content-type: image/avif.

Hey, I was talking about js, css and html. I didn’t compress images using brotli since they’re already compressed.

For main.bundle.js I see

Content-Encoding: br

in Chrome and Safari, but not Firefox.

But as Firefox can use it I cannot explain the missing header.

Firefox does show this header for main.bundle.css though.

But using compression had no impact on the speed of the website. How come? Everything seems fine: I generated br version of all js/css assets and html files. Compressing the js and css results in files 5 times smaller than the original.

We automatically brotli or gzip request most suitable asset types for you if the browser requests it; no additional config required or possible.

As per Zip deploy: Brotli and deletes; questions not problems - #2 by fool

Using curl as shown in Bundle file not served GZIP - #4 by luke the gzip version is 45917 and brotli version 44308—or about 4% smaller—as opposed to 212343 uncompressed (all in bytes.) Given the compressed size is 44/45KB and only 4% difference between the two, I see this having a negligible impact on load performance.