What SEO for ChatGPT covers, and what it leaves to others
SEO for ChatGPT has two halves, and this page is about the first: being findable. Before ChatGPT can name your business, it has to be able to reach your pages, find them in an index and retrieve them for a question. Those steps are technical, cheap and mostly binary. The second half, being the page worth citing once found, is about content and reputation. We cover that in our guide to how to get cited by ChatGPT.
If you would rather not touch any of this yourself, getting help is a reasonable choice. Our page on hiring help for AI search explains what a provider should handle. But the technical basics are an afternoon of work, and knowing them helps you check anyone you hire.
A note on the name. Plenty of people search for ChatGPT SEO, and the phrase pulls in two topics. This page is about getting your own site found by ChatGPT. If you are looking for how to use ChatGPT as a helper for ordinary SEO work, that is a different job, and we cover it in using ChatGPT for SEO.
Let OpenAI's search crawler in, get your pages into Bing's index, tell search engines when you publish, and make sure your main text shows up without scripts.
How ChatGPT search finds pages
ChatGPT answers many questions from what it already learned in training. When a question calls for current or specific information, such as a local business, a product or a recent change, it can search the web and build its answer from what it finds.
For that web search, it does not rely only on one source. OpenAI has said it uses search partners alongside its own crawler, and Bing's index is widely reported to be a major part of that. You cannot see the exact mix from outside, and it can change. The practical result is stable enough to act on: pages that are well represented in Bing's index and reachable by OpenAI's crawler are the ones that can show up.
This has a surprising consequence. Many site owners have put years into Google and ignored Bing entirely. A page can rank on page one of Google and be missing from Bing. If so, you have a gap that Google-only habits hide.
There are three steps to check, in order.
- Can OpenAI's crawlers read the page? This is a question about your robots.txt, your host and how the page renders.
- Is the page in the index? Bing Webmaster Tools tells you, and IndexNow speeds it up.
- Does the page match the question? That is content, and it belongs in the other guide.
The three OpenAI crawlers and what each one does
OpenAI documents separate crawlers for separate jobs, each with its own name. At the time of writing they are:
OAI-SearchBot collects pages for ChatGPT's search results. If you want to be shown and linked in those answers, this one needs access.
GPTBot collects pages that may be used to train OpenAI's models. Allowing or blocking it is a separate decision from search, and OpenAI states that blocking it does not affect whether you appear in search. Our entry on GPTBot explains it briefly.
ChatGPT-User fetches a page when a person using ChatGPT asks it to look at a link or run a lookup. It acts for a human at that moment.
Check OpenAI's own crawler documentation for the current list, because names and behavior change.
The mistake to avoid is blocking everything that has OpenAI in its name because you do not want your writing used for training. That choice is yours to make. But if you block OAI-SearchBot you give up the search listing as well, and a business that wants to be recommended should think about that trade.
To see what your file says, add /robots.txt to your web address. A rule that allows the search crawler looks like a User-agent line naming it, followed by an Allow line for the whole site. A rule that blocks it has Disallow with a single slash. Rules for one bot do not change the rules for another, which is why a broad block written years ago can still be in force. Our entry on robots.txt covers the syntax.
Bing Webmaster Tools and IndexNow
This is the step most owners skip, and it is often the most useful.
Set up Bing Webmaster Tools
Bing Webmaster Tools is Microsoft's free report on how Bing sees your site, much like Google Search Console. Create an account, add your site and verify ownership. If your site is already verified in Google Search Console, Bing offers a quick import. Then:
- Submit your sitemap, which is the list of your pages that your platform generates.
- Check how many pages Bing has indexed against how many you have. A big gap is worth investigating.
- Look at the crawl and indexing reports for errors. Pages Bing cannot reach are pages ChatGPT's search may not find through it.
Use IndexNow
IndexNow is an open protocol that lets a site tell participating search engines the moment a page is created, changed or deleted. Bing takes part. Instead of waiting for a crawler to come back, you send a short notice with the address of the changed page.
Many platforms and plugins support it already. WordPress has plugins that send a notice on publish. Some hosting and content delivery services offer it as a setting. If your platform has no support, you can send the notice yourself with a simple web request, which Bing's documentation describes.
IndexNow does not guarantee that a page will be indexed, or that it will be used in an answer. It shortens the delay between publishing and being known. For a business that publishes regularly, or changes prices and hours, that matters.
Rendering, speed and the other basics
Once crawlers can reach you and an index knows about you, check that your pages are readable by a machine.
Main text without scripts. Some crawlers read the raw page and do not run JavaScript. If your key text only appears after scripts load, they may see an empty shell. View the page source in your browser and search for a sentence from your main content. If it is not there, ask your developer or platform about server-side rendering. The problem is known as JavaScript SEO.
Clean titles and descriptions. Each page needs a title that says what it is, a description and one clear main heading. These are the labels an index uses to match your page to a question.
A working sitemap and sensible canonicals. The sitemap lists your pages. The canonical tag tells engines which address is the main version when several lead to the same content.
Speed and stability. A page that times out for a crawler is skipped. Keep the server responsive.
No accidental noindex. A noindex tag left over from a staging site is a classic way a whole site disappears. Check a few key pages.
Structured data. Markup describing your business, products or questions helps machines read the page. It is not a magic switch, but it removes guesswork. Look up schema markup if the term is new.
What about an llms.txt file, a proposed plain-text guide for language models? There is no strong evidence either way that major assistants use it, so treat it as optional and cheap, not as a fix.
An afternoon's setup, in order
- Read your robots.txt. Confirm OAI-SearchBot is allowed. Decide on GPTBot deliberately.
- Test your host. Confirm a bot gets a normal response, not a challenge page. Your host's support or logs can tell you.
- Add your site to Bing Webmaster Tools. Verify, submit the sitemap, compare indexed pages to real pages.
- Turn on IndexNow through your platform or a plugin.
- View source on five key pages. Check the main text is present and there is no noindex.
- Ask ChatGPT. Put five of your buyers' questions to it, and note who is named. Do the same next month.
That last step connects setup to results. If you are findable and still not named, the cause is no longer technical. It is that the pages are not the ones assistants pick, or that sources outside your site do not vouch for you. Our page on why your brand is not in ChatGPT walks through that diagnosis.
None of this guarantees a mention. It removes the reasons you could not be one.
Where to find these settings on common platforms
The steps differ depending on how your site is built. These are general pointers, and platforms change, so confirm in their current help pages.
Shopify. Shopify generates the robots.txt file for you, and it can be changed through a template in the theme. Most stores never need to touch it. Check yours by visiting /robots.txt and confirm no line blocks OpenAI's search crawler. Shopify also creates a sitemap automatically, so you submit that address to Bing Webmaster Tools.
WordPress and WooCommerce. A robots.txt file is either generated by WordPress or created by an SEO plugin, and some security plugins add their own bot rules. If a block appears that you did not write, search your plugin settings for bot or crawler controls. Many SEO plugins also include an IndexNow option.
Wix. Wix has built-in SEO settings, including a way to view and edit the robots.txt file in its SEO tools. It also provides a connection to Bing Webmaster Tools.
On any platform, remember that the host or a content delivery layer in front of the site may add its own bot rules. If robots.txt looks fine but crawlers still cannot get in, ask your host whether bot protection is turned on and how it treats search crawlers.
Our overview of Shopify ChatGPT optimization goes into the store-specific side in more detail.
Checking whether the crawlers actually visit
Allowing a crawler and being visited by it are different things. If you want proof, look in your server's access logs.
An access log is a record of every request made to your site. Each line includes a label identifying the visitor, called the user agent. Search the log for the crawler names above. If OAI-SearchBot appears with successful responses, it is reading your pages. If it never appears, either it has not come yet, or something is stopping it.
Look at the response code beside each visit. A normal reply means the page was served. Replies in the error range mean it was turned away. A pattern of refusals for bots, while normal browsers get through, points to a firewall or bot protection rule.
Not every host makes logs easy to read. If yours does not, ask support to search for the names for you. Some hosts and content delivery services also have a bot analytics screen that groups visits by crawler.
One caution. Anyone can put a crawler's name in a visit label, so a log line alone is not proof that the visitor was really OpenAI. OpenAI publishes the address ranges its crawlers use, and a security-minded host can check against them. For ordinary purposes, the log tells you enough.
If crawlers visit and your pages still do not appear when you ask, the cause has moved from access to quality, which is the territory of getting cited by ChatGPT.
How RunOctopus handles this
Findability is the floor. RunOctopus builds on it by writing and publishing the pages ChatGPT can then pick.
RunOctopus is a done-for-you service. Its AI, Ollie, researches your business, then writes and publishes pages directly on your own website. A team plans, writes, reviews and publishes alongside Ollie. It works with Shopify, WordPress, WooCommerce and Wix.
There is nothing to install, nothing to approve, nothing to upload and nothing to schedule. You do not approve each page. Pages go live on their own. In the first week you get 8 long-form articles, 6 pages and a custom interactive tool built for your niche. After that it is 175+ pieces of content a month, written, published and linked. You also get reports on where ChatGPT, Perplexity, Claude and Gemini name your business.
There is a 90-day promise: more visibility on Google and in AI answers within 90 days, measured against the 90 days before you joined, or every dollar back. Paste your address into the free site read and Ollie reads the site and shows the plan. No account, no call.