Do Steam and itch.io block AI crawlers in robots.txt?
No. On 2026-11-09 we read both robots.txt files: both block only function pages such as checkout, downloads, embeds and site search for all crawlers (User-agent: *), game detail pages are not blocked, and no AI crawler is named.
| Site | Paths blocked for all crawlers in robots.txt | Game detail pages | Sitemap |
|---|---|---|---|
| store.steampowered.com | /share/ · /news/externalpost/ · /widget/ · /email/ · /account/ackgift/ · /login/?*guestpasskey= · /join/?*redir= · /account/emailoptout/?*token= · /account/notificationsettings/?*token= | /app/… Allow | Not in robots.txt |
| itch.io | /embed/ · /embed-upload/ · /search · /checkout/ · /game/download/ · /bundle/download/ · /register-for-purchase/ · /email-feedback/ | Allow | itch.io/sitemap.xml |
Note that itch.io blocks /embed/ and /embed-upload/: the embedded player pages of browser games will not be read by crawlers that obey robots.txt, so the text on the game page is all they can read.
What can AI crawlers read on a Steam game page?
Quite a lot. We opened one Steam game page (Hollow Knight): the HTML is about 156 KB, each game has its own title and og:title, an og:description summary and a cover image, the price and rating are in the page as itemprop microdata, the game description and detail text are directly in the HTML, and there is no JSON-LD. Requests with the User-Agent strings of OAI-SearchBot and ClaudeBot both returned HTTP 200; these are simulated requests, not what real crawlers fetched.
| What we checked | Result (2026-11-09) |
|---|---|
| title and og:title | Hollow Knight on Steam |
| Description | og:description is the game's own summary; the detail text is also in the HTML |
| Price and rating | itemprop: offers, price, priceCurrency, aggregateRating, ratingValue, reviewCount |
| Structured data | No JSON-LD |
| Whether AI cites these pages | Not verified |
So what deserves care on a Steam page is the description: say the gameplay, genre, platform and languages in the first one or two sentences. This is the part that both AI and players read first.
We could not read the HTML of an itch.io game page: requests with simulated crawler User-Agents returned 403 or 429, and we stopped there without trying again. Its title, description and structured data are not verified.
What should a game website contain so AI can cite it?
The website is the only page you fully control and can put structured data on. You cannot change the layout of store pages, so the website should add what the store does not say clearly and stay consistent with the store description.
| Content | How to write it | Why |
|---|---|---|
| First paragraph | One-sentence definition: genre, core gameplay, platform, release date. For example: "Example Quest is a puzzle platformer for PC and browser, released in October 2026." | This one sentence is what AI needs when answering "which games of genre X are there" |
| Store links and price | Write Steam, itch.io and other links as text links, and the price as a number with currency | A price that exists only in an image cannot be read by crawlers |
| System requirements and supported languages | List minimum specs, operating systems, supported languages and controller support in a table | Players often ask for this, and store page fields do not always cover it |
| Gameplay and controls | Write it as text and tables, not only in the trailer | Crawlers cannot read what is in a video |
| Press kit page | Game summary, release date, developer name, contact, screenshot downloads, trailer links | Media cite it when writing reviews and recommendation articles, and those articles are one of AI's sources |
| Devlog | Write the date and what changed for each update | Gives the page a clear update time |
{
"@context": "https://schema.org",
"@type": "VideoGame",
"name": "Example Quest",
"description": "Same sentence as the first paragraph of this page.",
"genre": "Puzzle platformer",
"gamePlatform": ["PC", "Web browser"],
"applicationCategory": "Game",
"operatingSystem": "Windows, macOS, Linux",
"datePublished": "2026-10-01",
"url": "https://example.com/example-quest/",
"offers": { "@type": "Offer", "price": "9.99", "priceCurrency": "USD", "url": "https://store.steampowered.com/app/APP_ID/" }
}
Above is an example of the schema.org VideoGame type; the URL and name are placeholders. For how to write structured data and its limits see Does structured data help AI search?; for comparison pages see How to write comparison pages.
What does a browser game (HTML5) need on top?
The game itself is usually canvas or WebGL, and crawlers cannot read any text inside it, so add a readable description page around the game. For whether crawlers run scripts see Do AI crawlers render JavaScript?.
1. Write a paragraph above or below the game
Gameplay, goal, controls, how long a round takes, supported devices.
2. Host the game on your own domain
If the game exists only in an embed page on a third-party platform, crawlers that follow robots.txt cannot read that page (for example itch.io blocks /embed/). Put a text description on your own domain and then link to the platform page.
3. Write a "how to play" page and an update log
One question per page, answered directly in the opening.
4. Make titles unique
Give each game page its own title and description instead of one for the whole site.
How do you check it afterwards?
| Item | How to check |
|---|---|
| Your site does not block AI crawlers | Use the AI crawler check tool; for rules see the robots.txt guide |
| Text in the HTML | curl the homepage and search for a sentence of body text |
| Store description and website first paragraph mean the same | Read both once |
| Ask AI a month later | Ask "which indie games of genre X are there" and see whether you are mentioned and which URLs are cited |
Tool: AI crawler check; rules: How do you write robots.txt for AI crawlers?.
Outside voices also affect recommendations: for trailers see Do YouTube videos appear in AI answers?, for player discussion see Why does AI often cite Reddit?, for media reviews see Do review sites help AI recommendations?. For store page questions on mobile games see How do I get my app recommended by ChatGPT?.
If you make Unity assets or tools, see How do AI search tools find assets on the Unity Asset Store? Asset pages tested, plus the official rules for the Asset Store's robots.txt, sitemap, asset page test and the official Submission Guidelines.
If you make Minecraft mods, resource packs or server plugins, Modrinth and CurseForge checked and officially required are in How do Minecraft mods get found by AI search?.
FAQ
Will ChatGPT cite Steam pages?
We found no official statement and no actual citation evidence, so this is not verified. What we can confirm: the Steam robots.txt does not block game pages, and the page HTML has the game's title, description, price and rating (tested 2026-11-09).
Can AI read itch.io game pages?
The itch.io robots.txt does not block game pages, but our simulated crawler requests for a game page got 403 or 429 and we did not read the page content, so this is not verified.
Is a Steam or itch.io page enough without a website?
You can start by writing the store description well, but the website is the only place for structured data, comparison pages and press kit material. A one-page site is enough: definition, platforms, price, system requirements, store links and contact.
Does a browser game need a text description?
Yes. The game itself is canvas or WebGL and crawlers cannot read text inside it. Write the gameplay, controls and update log around the game.
Which structured data type should I use?
schema.org has a VideoGame type, where you can write genre, platform, operating system, release date and price. Fields must match the body text. Which AIs read it is not stated officially.
Sources
- Steam · store.steampowered.com/robots.txt (Read 2026-11-09)
- itch.io · itch.io/robots.txt (Read 2026-11-09)
- Steam · 示例游戏页(用来查 HTML 里有什么) (Read 2026-11-09)
- schema.org · VideoGame (checked 2026-11-09)
- kagentx evaluations: the data on the case study pages; each example gives its date.