Submit a list of domains and get back what each site is built with - CMS, ecommerce platform, analytics, CDN, hosting and web server, plus a single roll-up column listing everything detected. Newer runs add BuiltWith's own category breakdown on top.
This is the thing to know before you write an importer, because the column count is not stable across runs.
Every run returns nine lower-case columns: domain, then cms, ecommerce, javascript, analytics, cdn, marketing and server, and finally technologies - a comma-joined roll-up of what was detected. That set is the dependable core.
Newer runs return those nine plus twenty-five more: query, and twenty-four Title-Case columns carrying BuiltWith's own category names - Content Management System, Web Hosting Providers, SSL Certificates, Name Server and the rest. Thirty-four columns in total. The data dictionary below lists both sets and marks which is which.
Write the importer against the nine, and treat the twenty-five as present-if-present. A reader that indexes columns by position, or that assumes a fixed header, will break the first time it meets the other shape.
A domain is the entire input. There are no filters to configure.
Both forms work - a bare domain like stripe.com or a full URL like https://shopify.com.
Measured from 9 real exports covering 12 rows and 9 distinct domains. The first nine appear in every run; the rest appeared only in the two most recent ones.
Shopify.Shopify.Next.js and jQuery.Google Analytics and Google Tag Manager.Cloudflare and CloudFront.Intercom.nginx, cloudflare and Vercel.analytics.cdn.cms.Google Workspace in every newer-shape row we hold.Next.js.en-US and en.Viewport Meta in every newer-shape row we hold.Klarna.Robots.txt. Note the underscore in the column name.HTTPS.nginx and Cloudflare.Google Search Console, Facebook.nginx and cloudflare.Intercom.ecommerce. Note the lower-case first letter in the column name.HTML5 Video. Note the spaces around the slash in the column name.Counts, not percentages. The newer 34-column shape appears in only 4 of the 12 rows we hold, which is far too small a sample to express as a fill rate - so this table says "empty in all four" where that is what we measured, and claims nothing about how those columns behave at volume. The nine lower-case columns are on firmer ground: across all 12 rows, domain arrived every time, technologies on 10, server on 9, and cms, ecommerce and marketing on 3 each. Run a handful of your own domains on the free tier before you build around any single column.
Each of these came out of reading the exports rather than the documentation, which is why they are worth the space.
Two of the twelve rows we hold came back with every field empty except domain - including technologies. That is a real outcome, not an error: the site simply matched nothing. Treat an empty row as a normal result rather than a failed lookup, and your row count will keep matching your input.
The same domain run twice returned the same four name servers in two different orders inside Name Server. Split on the comma and compare as a set - diffing that field as a string will report a change that did not happen.
In one export Copyright came back as hds-text--md">© - a fragment of an HTML class attribute that happened to sit next to the symbol, not a copyright statement. Do not display this field without checking it, and do not parse a company name out of it.
cms and Content Management System describe the same detection, as do cdn and Content Delivery Network. The values are not always spelled the same way: one row carried server as cloudflare and Web Hosting Providers as Cloudflare. Normalise case before you join or group on them.
Where knowing what a site runs on changes what you do next.
If your product plugs into a specific platform, the platform is the qualifier. Running a domain list and filtering on cms, ecommerce or analytics separates the accounts where your integration works today from the ones where it would be a migration - two very different conversations, and worth knowing before the first email.
Run every domain in a sector and count the values in cms, cdn and server. The result is a picture of what that market has standardised on, drawn from the sites themselves rather than from vendor case studies - and the outliers are usually the interesting part.
Before committing engineering time to a partner integration, it helps to know what is on the other side. One row tells you the hosting, CDN, server and analytics a site runs, which is often enough to spot the mismatch early - or to confirm there isn't one.
No subscription, no minimum, no per-seat licence. Your first 500 rows are on us - after that, pay-as-you-go.
Every new account, one-time. No credit card required. Worth spending a few of them on domains whose stack you already know, so you can see which columns fill for the kind of sites you care about.
Roughly $2 per 1,000 domains, one row per domain - exact to budget, because a domain that detects nothing still returns exactly one row.
Volume pricing, SLAs, dedicated workers and bespoke onboarding for teams re-checking a large domain set on a schedule. Tell us your numbers and we'll quote.
LIVESCRAPER10 at checkout.This tells you what a site runs on. These two tell you who owns it and how it performs.
Twenty firmographic columns from the same input - name, industry token, headcount, founded year, revenue, address and socials. Stack plus firmographics qualifies an account far better than either alone.
Rank and engagement estimates for the same domain - global, category and country rank with their movement. What a site is built with, next to how it performs relative to its peers.
One of the lighter questions in the catalogue, and it is worth saying why rather than simply asserting it.
What a site is built with is published by the site itself. A server header, a script tag, a viewport meta tag and a DNS record are all things a browser reads on every visit - the detection here is the same observation, made deliberately and written down. There is no personal data in these columns: they describe software and infrastructure, not people.
Two honest caveats. The Copyright column lifts a line off the page, and on a one-person site that line can name an individual - treat those rows with the same care you would give contact-level data. And Name Server and Email Hosting Providers describe a domain's infrastructure, which is public but is also the kind of detail worth handling thoughtfully if you are publishing rather than analysing.
On our side: publicly available sources only, nothing behind a login, no third-party trackers on the data layer, and your exports auto-delete after 30 days.
The questions we hear most. Anything else? Talk to us - humans, not bots, write the answers.
domain, cms, ecommerce, javascript, analytics, cdn, marketing, server and technologies. Newer runs add query plus twenty-four Title-Case category columns - thirty-four in total. The data dictionary on this page lists all of them and marks which set each belongs to.technologies. Two of the twelve rows we hold came back that way. It is a real result rather than an error, and it means your output row count keeps matching your input.technologies is a comma-joined roll-up of what was detected. For anything programmatic, parse the individual columns. The roll-up's contents depend on which shape the run returned, so the same domain can produce a different string on a later run even when nothing about the site changed.Paste your domains and get the CMS, ecommerce platform, analytics, CDN and server behind each one. Your first 500 rows are free - enough to check which columns fill for the kind of sites you care about.
Knowing what a website is built with changes how you approach it. A prospect already running the ecommerce platform your product integrates with is a different conversation from one that would have to migrate first; a competitor that has standardised on a particular CDN and analytics stack tells you something about how the market it sits in operates. The BuiltWith Scraper collects that at scale. Submit a list of domains - a bare domain like stripe.com or a full URL like https://shopify.com both work - and each comes back as a single row describing the technologies detected on that site.
Every run returns nine columns: the domain itself, then the CMS, ecommerce platform, JavaScript framework or library, analytics and tag management, CDN, marketing widgets and web server, and finally a comma-joined roll-up of everything found. Newer runs return those nine plus twenty-five more - the submitted query echoed back, and twenty-four columns carrying BuiltWith's own category names, among them Content Management System, Email Hosting Providers, Frameworks, Language, Mobile, Name Server, Payment, SSL Certificates, Web Hosting Providers, Web Master Registration, Widgets and Audio / Video Media. That two-shape behaviour is the single most important thing to understand before writing an importer: code against the nine that are always there and treat the rest as optional, because a reader that assumes a fixed header will break the first time it meets the other shape.
Four details are worth knowing in advance, all of them observed in real exports rather than taken from documentation. A domain where nothing is detected still produces a row, with only the domain column filled - that is an ordinary outcome, not a failure, and it keeps output row counts matching input. The name servers in a single field come back in a different order between runs of the same domain, so that column should be split on the comma and compared as a set rather than diffed as a string. The Copyright column is not dependable: in one export it returned a fragment of an HTML class attribute that happened to neighbour the copyright symbol, so it should not be displayed unchecked or parsed for a company name. And the lower-case and Title-Case column sets overlap, describing the same detections with occasionally different capitalisation of the values, which matters if you group or join on them.
On the legal side this is among the lighter services in the catalogue, for a straightforward reason: what a site is built with is information the site publishes to every visitor. A server header, a script tag and a DNS record are read by any browser on any visit; recording them deliberately is the same observation written down. None of the columns describe a person. The two caveats worth stating are that the Copyright column lifts a line from the page and can name an individual on a one-person site, and that infrastructure detail like name servers and mail hosting, while public, deserves thought if you intend to publish it rather than analyse it. Nothing behind a login is touched, no third-party trackers run on the data layer, and exports auto-delete after 30 days. Start free: your first 500 rows cost nothing and need no credit card.