Download websites that use Wagtail

About Wagtail

Some content systems are chosen by the person who will write the content. Wagtail is chosen by the person who will deploy it, and almost everything interesting about the data follows from that distinction.

There is no one click path to it. Getting a site running means a Python environment, a framework install, a database, a deployment story and somebody who is comfortable with all four. Torchbox released it under the BSD in 2014 as a content layer for developers who wanted structured, block based editing without giving up control of the data model. Nobody arrives at it by accident and nobody arrives at it through a hosting panel promotion.

So the presence of this system on a domain is a statement about the organisation behind the domain, not about the website. It says there is engineering capacity somewhere in the picture, in house or contracted, and that somebody made a build versus buy decision and chose build. The population skews toward universities, public bodies, health and research organisations, media publishers, charities and mid market companies with a technical team. As of September 2026 we see it on 42,974 domains, which is a small number next to the mass market platforms and a much richer one per row.

For anyone selling into technical organisations, that is the whole value. A list of domains running this system is a pre qualified list of employers with a Python practice, of buyers for developer tooling, of accounts worth approaching about hosting, observability, accessibility auditing or security review. Recruiters read the same list as a hiring signal. None of those readers care about the CMS itself. They care about what running it implies about the people behind it.

Geography is the part that resists a tidy explanation. Germany accounts for 26,100 of the domains we detect and the .de zone leads the split, which does not match where the system was built or where its best known deployments are. The honest reading is that we are seeing hosting geography rather than organisational geography, plus the clustering effect of a few regional agencies and a hosting community that adopted it early. Treat the country field as where the site is served from, not where the organisation sits.

We publish no version for it, and the reason is consistent with everything else here. The application does not advertise its release by default, and the audience that installs it is not the audience that leaves defaults in place. Where other platforms hand us a version string for free, this one has to be inferred, so we would rather report nothing than report a guess.

The same caution applies to the count. Detection rests on admin paths, static asset fingerprints and response behaviour, all of which a reverse proxy or an aggressive CDN configuration can flatten. The sites most likely to be missed are the ones run by the most capable teams, so the population we publish is real but incomplete, and incomplete in the direction that matters. Read it as a floor and cross reference the rest of the row before drawing conclusions.