Not open yet — signups haven't started. Join the list for the day they do.
Home/Features/Global CDN

Global CDN

Fast wherever your reader is

Your site is kept close to the people reading it, all over the world, so a page opens quickly in Manchester and Manila alike — and holds up on the day something takes off.

Frankfurt edge
14ms · cache HIT
Live
Singapore edge
19ms · cache HIT
Live
São Paulo edge
22ms · cache HIT
Live

Every site here sits behind a content delivery network — which is just a lot of servers spread around the world, each holding a copy of your files. A visitor's request gets answered by whichever one of those copies is physically closest to them, instead of every request worldwide converging on one machine in one location. That's edge caching in a sentence: keep a copy near the reader, not just near the writer.

What a request actually does

It never has to go far

A visitor's browser talks to the nearest edge location, not to one server on the other side of the map. If that edge already has the file cached, the answer comes straight back — nothing gets assembled on the fly, and nothing has to wake up first. That's the entire difference between a cached page and one served from a single origin: less distance to travel, and nothing standing between the file and the wire.

GET /index.html
visitor → nearest edge location
cache check… HIT
200 OK · served without contacting origin

What edge caching absorbs

The load that never reaches one machine

250 GB
Traffic included
every byte served counts once, from the edge
0
Servers to configure
edge caching is on by default
1
Origin, many edges
a traffic spike is absorbed by cached copies
99%
Uptime target
the same target at every edge location

250 GB is 250,000 MB — a 2 MB page could be viewed roughly 125,000 times before you'd approach it, and because cached responses come from the edge instead of round-tripping to one server, that traffic doesn't concentrate anywhere it could bottleneck. Go over, and we email you before anything changes; there's no automatic overage charge.

What it means for speed

Nothing to wake up, nowhere far to reach

A page waiting at the edge skips the two things that usually slow a site down. Standard on every site here, and it works the same after a restore from daily backups.

Nothing has to wake up

The page is already sitting there, ready. It gets handed straight to the browser that asked for it.

Nothing gets rebuilt

The page was finished when you published it. What's waiting at the edge is exactly what your reader gets.

Same page, every time

It looks and loads the same whoever asks, whenever they ask, and from wherever they happen to be.

The moments it isn't cached yet

What happens when the edge doesn't have it

Caching only works once something's actually been cached — here are the two moments where that isn't true yet.

  1. First request, or a cache miss

    The first visitor to a page — or one after the cached copy expired — triggers a quick trip back to origin to fetch the current file. That answer is then cached at the edge for the next visitor, so misses are rare and get rarer as a page settles.

  2. You push a new deploy

    Publish a change and the edge copies of anything that changed are invalidated automatically, so the next request pulls the new version instead of serving a stale one for hours. You don't flip a cache-clearing switch — it's just part of what a deploy does.

A CDN isn't a plan you upgrade into here — it's how every site is served, always.

Plenty of companies sell edge delivery as a paid tier, or leave you to wire up an account elsewhere and work out the caching rules yourself. There's nothing here to configure and no setting to turn on — every site, from the very first publish, goes out across the same edge network as every other one. Speed isn't the upgrade; it's the floor.

Worth knowing

Quick answers on the CDN

Does the CDN cost extra?

No — it's how every site is served by default, included at every price point, not a separate product.

What happens on a cache miss?

The edge makes a quick round trip back to origin, serves that answer, and caches it for the next request — misses get rarer as a page settles.

How fast does a deploy reach the edge?

Changed files are invalidated at the edge as part of publishing, so the next request after a deploy gets the new version rather than a stale cached one.

Push a change, the edge catches up

Works the same whether you deploy by Git or the CLI. See the full numbers on pricing.

Join the list