Written by James Britt
GitHub, Shopify and Airbnb turn up in almost every list of websites made with Ruby. Familiar names are useful, but they only answer part of the question. How many sites use Ruby, and what does it mean when a crawler says it has found one?
On 16 September 2026, WebAtla’s Ruby technology page reported 523,100 detected sites. That is a useful starting point, provided we are clear about what has been counted. A technology label is evidence from a crawler, not access to a company’s source code.
How many websites are made with Ruby?
What counts as a website made with Ruby? A Rails application handling visitors’ requests clearly qualifies. So might a static site produced with Ruby tools, or a single Ruby component inside a much larger system. These are quite different uses of the language.
A crawler cannot always tell those cases apart. It also cannot see every application behind a login or identify a backend that leaves no useful public clues. Missing detections matter, but so do mistaken classifications and old signals.
That makes the count an estimate of detectable usage. Calling it a guaranteed minimum number of live Ruby backends would go too far.
Static sites are a good example. Jekyll uses Ruby during the build process and produces static files. The resulting website can be served without a Ruby application running for each visitor. It was built with Ruby tooling, but its hosting requirements differ from those of a Rails application. Jekyll’s documentation explains how that process works.
Counts change between crawls, too. Sites go offline, teams switch stacks and detection rules improve. The total alone cannot tell you which explanation applies. You need comparable snapshots and some investigation of the domains that entered or left the list.
Ruby among the backend languages
WebAtla’s programming-language category showed the following counts when checked on 16 September 2026:
| Language | Detected sites |
|---|---|
| PHP | 37,868,303 |
| Java | 965,209 |
| Python | 868,795 |
| Ruby | 523,100 |
| Perl | 208,712 |
These are figures from one provider’s programming-language index, rather than a census of applications.
PHP has by far the largest count in this selection. Ruby falls below Java and Python and above Perl. Keep the comparison within those limits: different languages leave different clues, and a provider may infer one technology from another.
Nor does one domain equal one development team. Several domains might belong to the same business, while one domain can front many applications. You cannot turn these totals into developer counts, market revenue or the number of new projects started this year.
The comparison is useful for understanding the dataset. Choosing a language for your next application calls for different evidence: the work it needs to do, the skills of your team and the software you expect to maintain.
Notable websites made with Ruby
For individual websites, an engineering team’s own account is usually more useful than a technology label.
GitHub has published a detailed account of its Rails monolith and how its engineers test and upgrade Ruby and Rails. That gives readers a view of the application itself and the work required to keep it running. GitHub’s engineering article is a stronger source than an undated company list.
Shopify has also described its Ruby on Rails codebase and its move towards a modular monolith. Its account explains why the company kept a shared application while introducing clearer boundaries between components. See Shopify’s engineering discussion.
WebAtla’s samples included GitLab, Goodreads and ThemeForest. Its August snapshot also listed Dribbble. Treat these as dated observations, subject to the detection limits above.
Historical examples need a date attached. For a company such as Kickstarter, an old Rails case study can establish past use; it cannot describe the complete production stack today.
For any company, the question is which application used Ruby and when. A checkout, an internal dashboard and a public marketing site can all run on different stacks.
Where Ruby sites are hosted
Using the country counts shown on WebAtla’s Ruby page, the breakdown below recalculates each share against the full 523,100-site total.
| Country by IP | Detected sites | Share of total |
|---|---|---|
| United States | 337,479 | 64.5% |
| Canada | 38,336 | 7.3% |
| France | 22,205 | 4.2% |
| Germany | 20,754 | 4.0% |
| Russia | 15,219 | 2.9% |
The US and Canada together account for about 71.8% of that total. These percentages use an explicit denominator; they differ from the country shares displayed by WebAtla.
“Country by IP” needs careful reading. It describes the location assigned to the detected IP address. A CDN or proxy may stand between that address and the application server. The field therefore does not establish where the business, its developers or its customers are based.
The same snapshot gives this domain-extension breakdown:
| TLD | Detected sites | Share of total |
|---|---|---|
| .com | 282,103 | 53.9% |
| .org | 30,422 | 5.8% |
| .net | 29,938 | 5.7% |
| .io | 23,210 | 4.4% |
| .ru | 12,483 | 2.4% |
Both tables use counts from WebAtla’s Ruby dataset page.
The .com lead is clear. Does the .io figure tell us that startups favour Ruby? We would need a comparable baseline and a way to identify those businesses before drawing that conclusion. A domain ending cannot answer it.
How this is measured — and how to pull it yourself
WebAtla describes a process that fetches pages, headers and scripts, matches technology signatures, and adds information such as IP country and ranking. Its Technologies dataset includes a date showing when each site’s technology information was last checked. WebAtla’s dataset documentation lists the fields.
That date matters. A fresh export does not necessarily mean every domain was checked that morning.
The provider offers downloadable JSON Lines data and a free sample. Check the package before using it: the Technologies dataset and the broader All data dataset contain different fields. DNS and WHOIS information should not be assumed to come with every technology export.
To analyse an export yourself, first record its date and filter for the Ruby label. Then look at individual results. Are you counting Rails applications, static sites generated with Ruby, or both? Write down that decision with your totals. Otherwise, someone repeating your work may count a different population.
Country and TLD filters can narrow a research list. Confirm individual companies before treating it as a list of Ruby employers or customers.
Frequently asked questions
How many websites are made with Ruby?
WebAtla reported 523,100 detected sites on 16 September 2026. That is a provider-specific count, not a verified total of every live Ruby backend.
What famous websites are made with Ruby?
GitHub and Shopify have published engineering accounts of their Rails applications. Other examples require a dated source and a clear distinction between historical use, crawler detection and the current production stack.
Is Ruby still used for web development in 2026?
Yes. Rails continues to be maintained, and its official site provides documentation and development news. In July 2026, the Rails team also introduced a library of real applications for developers to study.
How can I get a list of websites made with Ruby?
Use a technology dataset that supports Ruby filtering, inspect its sample and check the dates on its records. WebAtla offers JSON Lines exports; the fields available depend on the dataset selected.