Download websites that use Amazon CloudFront

About Amazon CloudFront

A CDN is one of the few technologies on a website that is genuinely bought. Somebody signed up, attached a billing account and pointed DNS at a vendor, and an invoice arrives every month whether or not anyone thinks about it again. Amazon CloudFront is the delivery arm of a much larger platform relationship, and detecting it on a domain usually means detecting the platform underneath.

The inference runs further than the edge itself. A site fronted by this service is, in most cases, a site with origin infrastructure at Amazon as well, with an engineer or an agency who works in that console daily, and with a procurement process that has already been through cloud billing and its approvals. As of September 2026 we detect the service on 1,871,782 sites, which is a population of organisations running real infrastructure rather than a shared hosting plan. The technology stands in for engineering maturity and for budget, and both are hard to observe any other way from outside.

United States accounts for 1,385,176 of the sites in that population, and .com carries the bulk of the domain mix. The concentration is partly a market fact and partly a billing one, since a global platform priced in dollars acquires customers unevenly. For a sales team the practical value often sits in the residual, which is the set of domains outside the concentration where the same infrastructure runs with far less competitive noise around it.

Detection is header and hostname based, and the limits deserve plain statement. We identify the service from response headers, asset hostnames and behaviour the edge exposes. A configuration that strips or rewrites those headers, or another proxy placed in front of the edge, can hide the service completely, and a site may also use it for images and video while serving its HTML from somewhere else entirely. A match means the service is involved in delivering that site, not that it delivers all of it, and certainly not that it is the only vendor in the path.

There is no version field for a managed service like this, and the blank is informative in itself. Edge platforms upgrade themselves and expose no per-site release, so there is nothing to fragment and nothing to date. The service is Proprietary and has been sold since 2008, so unlike an open source component its presence carries a price tag by definition. What the data cannot see is the size of that price. Traffic volume, feature tier and contract value stay invisible from outside, and any spend estimate built on presence alone is a guess wearing a suit.

The buyers of this list tend to be competitive or consultative. Rival edge and cloud vendors want a migration list. Consultancies and resellers want organisations already committed to a platform they specialise in. Security and monitoring vendors want to know whose traffic passes through which edge network. Researchers want the concentration itself, because the share of the web sitting behind a handful of edges is a resilience question rather than a marketing one. A lookup answers whether one site uses the service. The download answers how many, where, alongside what, and in which direction the population is moving, and only the second kind of answer supports a decision.