Put your site behind the network and every visitor is served by the nearest node, with the page already waiting in cache. Your own server stays where it is: nothing moves.
You create a Pull Zone pointing at your server's address. From then on the nodes fetch from it once and serve their own copy after that.
To send us the traffic you have two routes: a CNAME from your domain, or moving the DNS to us and flipping a switch. The HTTPS certificate arrives on its own, free, and renews on its own.
How long things stay in cache, with a default and rules by extension or path. You can tell a node to ignore what your server says, or to follow it.
What to do with query parameters: treat them as different pages or ignore them, so an ad campaign does not multiply your cache.
Whether to serve the stale copy when your server does not answer: the site stays up even while the origin is down.
Which nodes serve you. If you have a constraint on where data may sit, take the ones you do not want out: they leave the DNS and never even receive the site's configuration.
Every node compresses with Brotli, and with gzip for clients that do not support it. On a real page: 296 KB uncompressed, 60 KB with gzip, 38 KB with Brotli. One switch, nothing else to configure.
From the panel you purge one address, a folder or the whole zone. The request goes to every node and each one confirms when it is done: you do not have to take our word for it.
Bytes served, requests, cache ratio and how much your server was spared, split by node and by day. Plus the benchmark, which measures response time from every node before and after the CDN.