Guides / Guide
How to get your website cited by ChatGPT, Perplexity and Claude
What you control
- Access for the search crawlers. Each assistant has a crawler for search and a separate one for training. ChatGPT search uses OAI-SearchBot, Perplexity uses PerplexityBot and Claude uses Claude-SearchBot. Blocking a training crawler such as GPTBot or ClaudeBot does not block these. Open your robots.txt, then ask your server: a firewall or CDN can answer 403 to a crawler that robots.txt allows. The free AI crawler checker tests both, and /bots has the robots.txt lines to copy.
- Text in the HTML. Many crawlers do not run scripts. A page whose text is added by JavaScript reaches them as an empty shell. Open "View page source" and search for the sentence that answers the question; if it is not there, a crawler that does not run scripts cannot quote it.
- One page per question. A page that states the answer in its first lines, with the facts a buyer asks about (price, availability, who it is for), is easier to quote than a page that touches ten topics. Write the page for one question, in the words people ask it.
- Being in a search index. ChatGPT search draws on search-engine results from partners, and Bing is the one most often named. Add your site to Bing Webmaster Tools and Google Search Console and submit your sitemap, so the pages can be found at all.
What nobody controls
- Which pages an assistant names changes from one run of the same question to the next.
- OpenAI does not publish a citation report for site owners, so you see its crawlers' requests and visits from chatgpt.com in your server logs and the rest only by asking the questions yourself.
- A mention without a link may send nobody to your site.
- Files such as llms.txt are cheap to add, but no major assistant has confirmed that it reads them.
Be wary of anyone who sells a guaranteed citation. The part that can be fixed is access, text and structure, and that is the part to do first.
Blogs and local businesses
A blog helps when a post is the page that answers a question directly; more posts on the same topic do not add reasons to quote any of them. A local business gets the same three controls and one more habit: put your name, address, opening hours and what you sell in the HTML of a page of your own, so that the facts an assistant gives about you come from you.
Check where you stand
- Run the free AI crawler checker on your domain: it shows the robots.txt rule and the live response for each crawler.
- Ask the questions your buyers ask in each assistant, twice, and note which of your pages are linked: how to check by hand.
- For the whole site, CitedBy audits your pages up to your plan's page limit and gives the exact fix for each issue, from €9 paid once.
Questions people ask
How long until ChatGPT cites a new page?
We cannot give a number. The page must first be crawled and indexed, and answers change from run to run.
Does blocking GPTBot stop ChatGPT citing me?
Not by itself. GPTBot is the training crawler; OAI-SearchBot is the one used for search, and ChatGPT-User fetches a page when a person asks about it. Check each rule on its own.
Do I need an llms.txt file?
No major assistant has confirmed that it reads one. It is cheap to add and cheap to check, but access and text matter more: see does llms.txt work?.