Both frames render identically — that's the goal. The win is in the bytes and requests below, not the look. They're served live by mod_pagespeed 1.1 on demo-httpd-1.1.modpagespeed.com; the optimized frame applies only this filter. Right after a cache purge it may briefly match the original while the worker rewrites it — reload to see the result.

This filter needs configuration or content the public demo doesn't provide (e.g. an Analytics account id or a recognized JavaScript library), so there's no diff to show here — see the filter docs for what it does.