What CMS does this site use?
Paste a URL and check the CMS. Editors only change what the content management system exposes, so the CMS decides what is editable copy and what is locked in templates or code — and it decides which fixes are even possible. Judging which CMS a live site runs on means reading structural evidence — path prefixes, response headers, asset hosts — not hunting for a brand name inside marketing copy. A documentation page that mentions several vendors is not the same thing as a live install. This hub explains the outside signals, lists every CMS detector on this site, and separates a keyword hit from a confirmed stack.
How a CMS can be identified from the outside
Proprietary resource paths. Installs leave folders and URL prefixes that almost never appear on unrelated stacks. Examples include /wp-content/ for classic PHP blogs, /cdn.sanity.io for structured headless assets, and similar prefixes tied to one vendor's media pipeline. When those paths show up as real script, stylesheet or image URLs — not only as plain text in an article — they are strong structure.
Response headers and generator tags. Some platforms still emit x-powered-by values or a meta generator string that names the publishing engine and often a version. Headers can be stripped by a reverse proxy, so absence proves nothing; presence is one more structural vote when it agrees with path and host evidence.
Branded domains and API hosts. Media and API traffic often rides vendor hostnames such as images.prismic.io or cdn.shopify.com. Those hosts appear in src and href attributes of resources the browser must fetch. A blog post that merely discusses the same hostname is weaker than an actual resource request on the page you opened.
Page structure and draft or preview endpoints. Slice markup, data attributes, preview query parameters and draft toolbar scripts can reveal a CMS even when marketing copy never says the product name. Preview endpoints that leak into production are especially telling because they are part of the editing workflow, not something a competitor article would invent on someone else's domain.
Thirteen CMS detectors on this site
Each card below opens a dedicated checker. Paste a URL there to see match counts and the exact evidence strings. The roster covers every CMS checker shipped here today — headless, classic PHP, and hosted builders alike.
- Contentful— Headless entries and asset hosts that show up as space-scoped API and image URLs.
- Craft CMS— Twig-driven publishing with transform URLs and control-panel path fingerprints.
- Drupal— Module and theme directories under sites/default plus core library paths.
- Ghost— Publishing API and theme assets aimed at newsletters and long-form posts.
- HubSpot CMS— Marketing hub pages whose scripts and modules ride HubSpot content hosts.
- Joomla— PHP CMS marked by /templates/ folders and component query strings.
- Prismic— Slice-based headless CMS with images.prismic.io and repository API hosts.
- Sanity— Structured content served from cdn.sanity.io and Vision/GROQ tooling.
- Squarespace— All-in-one builder whose static CDN and site CSS follow a fixed layout grammar.
- Strapi— Self-hosted headless API with /uploads and admin path conventions.
- Webflow— Visual builder export marked by website-files hosts and data-wf attributes.
- Wix— Hosted site builder whose viewer scripts and wixstatic media paths stay consistent.
- WordPress— Classic PHP CMS fingerprinted via media, includes and REST route prefixes.
Why a keyword hit is not a confirmation
Brand strings appear in places that have nothing to do with the live stack. Vendor documentation sites discuss many products on one page. Job posts list tools the team hopes to hire for. Comparison articles name every platform in a category. In those cases a plain-text keyword can fire while no proprietary path, header or asset host is present.
This site therefore distinguishes suspected from confirmed. A suspected label means the brand token appeared without structural backing. A confirmed label means at least one structural pattern matched — a path, header, host or markup shape that ordinary prose does not produce. Prefer under-reporting over claiming an install that is only mentioned in copy.
What usually needs fixing once the CMS is known
Image pipelines. Most CMS installs resize or transform media through a vendor URL grammar. Oversized heroes, missing width and height, and transforms that still ship multi-megabyte files are common once you know which pipeline is in play.
Fonts and scripts. Themes and slices often pull webfonts and widgets from third party hosts. Knowing the CMS tells you whether those tags live in a theme file, a slice, a tag manager, or a site-builder app panel.
Caching and rendering strategy. Headless setups lean on CDN caching of API JSON and HTML shells; classic PHP stacks lean on page caches and object caches; hosted builders expose fewer knobs. The CMS category narrows which of those levers actually exist before anyone starts changing code.
How to check a website's CMS in three steps
Step one — view source, then search it. Open the page, press Ctrl+U (Cmd+Option+U on a Mac) and search the HTML for the vendor names you suspect. If a name only appears inside article text or a comment, it proves nothing. What counts is a name appearing in a src, href or data- attribute — that is a resource the browser was told to fetch.
Step two — read the headers. Open DevTools, switch to the Network tab, reload, and click the document request. Look for x-powered-by, x-generator, server, and any set-cookie names. A cookie called wp-settings or shopify_ is a stronger vote than a meta tag, because cookies come from the platform itself. Headers get stripped by CDNs, so a missing header proves nothing — treat it as one vote, not a verdict.
Step three — compare against a known install. The same platform leaves the same shape everywhere: the same asset host, the same URL prefix, the same editor endpoint. That is what makes a CMS check reliable. A single hit can be a false positive; a hit plus the matching host plus the matching path is a confirmation. If you want to skip all three steps, paste the URL into any checker on this page and read the evidence list it returns.
Signals left behind by common platforms
WordPress: /wp-content/ and /wp-includes/ asset paths, /wp-json/ REST routes, a wp-emoji-release script, a meta generator tag, and wp-* cookies. Shopify: cdn.shopify.com and cdn.shopifycdn.com assets, /cdn/shop/ paths, a Shopify.theme global in page scripts, and a _shopify_ cookie.
Wix: static.wixstatic.com and static.parastorage.com assets, a wix-image or wix-code script, and the viewer model JSON inlined in the page. Squarespace: static1.squarespace.com assets, a squarespace. prefixed script, and Static.SQUARESPACE_CONTEXT in the page. Webflow: website-files.com and website-images-assets.com assets, plus data-wf-* attributes on the html element.
Prismic: images.prismic.io and /prismic-assets/ media, plus slice markup and the prismic-toolbar or preview script. Contentful: images.ctfassets.net and assets.ctfassets.net media hosts. Sanity: cdn.sanity.io assets and a sanity project id referenced by the page script.
Ghost: a /ghost/ or /content/ path, ghost-* meta tags, and the Ghost admin SDK script. Drupal: /sites/default/files/ paths, Drupal.settings or drupalSettings in page script, and an X-Generator header. Joomla: /media/jui/ or /media/system/ assets and /components/com_ path segments. Craft CMS: a CRAFT_CSRF_TOKEN cookie and /cpresources/ asset paths. Strapi: an /uploads/ media path combined with an api. host or /api/ JSON endpoints.
CMS checker vs. the tools you have probably tried
Most "what CMS" answers online come from one of three shapes. Paid technology-lookup databases gate the useful rows behind a subscription and price per report. Hosted checker sites give you a verdict but not the reason, so a false positive is indistinguishable from a real match. Theme galleries infer the platform from a screenshot, which is a guess wearing a table.
This checker is built the other way round: every verdict lists the exact signature that matched — a script URL, a response header, a resource path, a DOM attribute — with the number of hits. If only a brand keyword appears and no structural evidence does, the result is marked likely rather than confirmed. That is the difference between a report you can paste into a client email and one you have to re-verify by hand.
It is free, there is no account, and the URL you paste is not stored: the request is analysed and the response returned. The same paste also reports front-end framework, hosting and CDN, third-party scripts and the plugins a CMS exposes — one check answers the stack question instead of five lookups.
Use a paid database for historical market share or a company's whole portfolio. Use a hosted checker when the platform is mainstream and you only need yes or no. Use this page when the answer has to survive someone asking how you know — and when the next question is what to fix rather than which CMS it is.
Start from the detector that matches the strongest structural clue you already see. If nothing is obvious, pick any CMS checker below and read the evidence list — a miss with a clear “likely” note is still useful, because it tells you the brand appeared without an install signature.
Optimization help is available
Tell me your URL and what you want fixed — you get a clear scope, price and timeline.
To:[email protected] | Cc:[email protected]