If you ran a marketplace, you would count your inventory and call that the size of your site. Products in, pages out.
We counted itch.io expecting exactly that, and the catalogue turned out to be the smaller part of it.
Two thirds of what itch.io publishes was made by its users doing something other than selling: writing about their work, listing other people's games, running competitions and arguing in forums.
Results at a glance
itch.io lists about 1.96 million games. It publishes about 3.9 million other pages that exist because somebody bothered.
The marketplace is a third of the marketplace.
The catalogue is the small half
Every sitemap file is named for what it holds, so the composition reads straight off the index.
Devlogs are developers writing about their own work in progress. Collections are lists a user made. Jam entries are games submitted to a competition somebody organised. None of it is inventory, and all of it is indexable.
A list one person made
The collections layer is the one worth staring at, because it is the least obvious million pages on the internet.
That page cost itch.io nothing. Somebody clicked "add to collection" six times, years ago, and stopped. It has been sitting there ever since, linking to six games with real titles and real genres.
Multiply it by 1,047,387 and you have a hand-built internal linking layer that no team could have produced and no competitor can copy, because copying it would mean acquiring the people who made it.
Why devlogs beat descriptions
The second-largest layer is stranger still: 1,205,533 devlog posts, more pages than there are collections.
A game's store page says what the game is. A devlog says what changed this week, what broke, and what the developer is trying next. It is the only part of the site written in prose by somebody with something to say.
For a marketplace, that is enormously valuable and almost free. The developer writes it because they want players to know, not because anybody asked for content.
This is the difference between user-generated content and this. Nobody on itch.io is producing pages for itch.io. They are talking to their own audience, on a page itch.io happens to own.
What a million curators earned
Domain rank 94, and 1,855,168 referring domains — the second-widest profile of anything we have measured, behind only Eventbrite.
There is an important caveat, and it is structural. itch.io gives every creator their own subdomain: the games in that collection live at addresses like daniellinssen.itch.io. Some portion of a "referring domains" count for a site built this way will be its own creators linking between their own pages. We cannot separate the two from the outside, so treat the figure as an upper bound.
What is not in doubt is the shape: developers link to their own itch.io page from everywhere, because it is where their game lives.
The cleanest speed score we have seen
Our Website Audit gave itch.io 88 out of 100 with zero critical issues, and a perfect 29 out of 29 on speed — the only full marks on performance in this entire series.
A site serving six million pages of user content is faster than the ad-funded publishers we have measured, because it does not carry their advertising stack.
Its faults were small and human: images missing alt text, a meta description that needs work, some outbound links pointing at dead pages. On a site where users write most of the content, dead outbound links are somebody else's rot arriving on your pages.
Copy this in an afternoon
The transferable idea is not "get user-generated content". It is narrower and much more useful.
1. Give people somewhere to put an opinion about your inventory. A list, a shelf, a wishlist, a comparison. Collections are itch.io's biggest surprise and the mechanic is a save button.
2. Let sellers talk to their own buyers on your domain. Devlogs work because developers have a reason to post that has nothing to do with you. Find the update your suppliers already send by email and give it a URL.
3. Make the container indexable from day one. A collection with six items and no updates for three years is still a page with six internal links. It only counts if it is crawlable.
4. Watch what your users break. Dead outbound links and missing alt text are the tax on this model. Budget for cleaning up after the content you did not write.
What you cannot copy
A community that was already there. itch.io's layers exist because game jams, devlogs and curated lists were part of indie game culture before itch.io existed. It built homes for behaviour people already had. If you add a collections feature to a marketplace where nobody wants to curate anything, you will get an empty feature and a million thin pages, which is worse than not building it.
And here is what we could not check:
The total backlink figure was not usable. Our tool reported a total that does not reconcile with its own dofollow and nofollow split for this domain — the two differ by more than twelvefold, where every other site we have measured differs by almost nothing. We have published the referring domain count and the follow split, and left the total out. That is a fault in our tool, not a finding about itch.io.
Referring domains may include itch.io's own subdomains, as described above. The figure is an upper bound.
Declared pages are not indexed pages. A three-year-old collection of six games is exactly the sort of page Google may decline to keep.
No traffic or revenue data. None was available and none is claimed.
The Backlink Analyzer separates referring domains from raw link counts, which is the difference between many sites mentioning you and one site linking a thousand times. If you are building a community layer, the domain count is the number that tells you whether it is working.
- The same check that returned 1,855,168 referring domains for itch.io
- Referring domains, anchors and the follow split, counted separately
- 3 checks a day, no signup, no card
Frequently asked questions
How many pages does itch.io have?
5,841,233 URLs across 122 sitemap files, counted on 25 August 2026 by fetching every file rather than sampling, with no errors. By type: games 1,955,552, devlog posts 1,205,533, collections 1,047,387, user profiles 895,953, jam entries 545,077, forum topics 109,576, jams 81,206 and browse pages 949.
So how much of it is the actual store?
About a third. Games are 33.5% of the total. Devlogs and collections together come to 38.5%, which is more than the catalogue, and neither was produced by itch.io.
What is its backlink profile?
Domain Rank 94 out of 100, with 1,855,168 referring domains, 27,056,451 dofollow and 18,817,664 nofollow links, 1,440,563 referring IPs and 1,340,616 referring subnets. We have deliberately not published a total backlink figure, because our tool's total did not reconcile with its own follow split for this domain.
Why might the referring domain count be inflated?
Because itch.io gives every creator a subdomain of the form name.itch.io. Depending on how a backlink index treats subdomains, links between creators' own pages can be counted as separate referring domains. We cannot tell from outside how much of the 1.86 million that accounts for, so we treat it as an upper bound rather than a clean figure.
What did the website audit find?
88 out of 100, with 0 critical issues, 2 warnings and 6 notices against 113 checks passed. On-page SEO 23 of 25, technical SEO 60 of 67, speed 29 of 29 and mobile 2 of 2. That perfect speed score is the only one in this series. The flagged items were images missing alt text, a meta description needing work, outbound links pointing at dead pages and links going through redirects.
How does this compare with others in this series?
Referring domains per page: calculator.net 143 across 222 pages, Healthline 14.4, GeeksforGeeks 0.6, itch.io 0.318 across 5,841,233, Eventbrite 0.148, PyPI 0.082, Levels.fyi 0.026. itch.io's ratio is the best of the multi-million-page sites we have measured, with the subdomain caveat above.
Is a collections feature worth building?
Only where people already want to curate. The signal to look for is whether your users are making lists somewhere else — in spreadsheets, in forum posts, in social threads. If they are, giving them a native place to do it converts existing behaviour into pages. If they are not, you are building an empty feature that generates thin URLs.
How is this different from ordinary user-generated content?
Most user-generated content asks people to write for the platform. These layers work because nobody is doing that. A developer posts a devlog to tell their own players what changed; a curator makes a collection to keep track of games they liked. The pages are a side effect of people pursuing their own reasons, which is why they keep coming without being managed.
The lesson here is quieter than most in this series. itch.io did not out-publish anybody. It noticed the things its users were already doing and made sure each of them produced a URL.
Your customers are probably doing something similar right now, somewhere you cannot index.
Run the check on your own domain and see how many separate sites actually link to you.
Measured on 25 August 2026. Page counts come from itch.io's own sitemap files; all 122 were fetched and counted rather than sampled, with no errors, but declared URLs are not indexed pages.
Link and audit figures are free runs of our own tools. The total backlink figure our tool returned for this domain did not reconcile with its own dofollow and nofollow split and has been omitted. No traffic or revenue data was available to us and none is claimed here.