How to Get Your Website Cited by ChatGPT and Perplexity
By the Techprime team · · 6 min read
Key takeaways
- Allow OAI-SearchBot and PerplexityBot. These are the crawlers that put a site into ChatGPT search and Perplexity results.
- Blocking GPTBot is a separate choice about model training. OpenAI says each setting is independent.
- Research suggests citations, quotations and statistics can raise visibility, with results varying by domain.
- Neither company publishes how it chooses sources, so nobody can guarantee a citation.
On this page (5)
To give your website the best chance of being cited by ChatGPT and Perplexity, allow their search crawlers in robots.txt, keep your content readable as plain text in the HTML, and publish specific, sourced information that other pages do not have. No one outside those companies can guarantee a citation, because neither publishes how it chooses sources.
Which crawlers should you allow?
The first step is simply being reachable. A security rule added years ago can block AI crawlers without anyone noticing. These are the agents that matter, according to each company's own documentation:
| Crawler | Used for | If you block it |
|---|---|---|
| OAI-SearchBot | Surfacing sites in ChatGPT search | Your site is kept out of ChatGPT search answers |
| GPTBot | Training OpenAI's foundation models | Your content should not be used for training. OpenAI does not say this affects search |
| ChatGPT-User | Visiting a page when a user asks | Not used to decide what appears in search |
| PerplexityBot | Surfacing and linking sites in Perplexity results | Your site may not be surfaced in Perplexity |
| Perplexity-User | Fetching a page a user asked about | Perplexity says it generally ignores robots.txt, since a user requested it |
OAI-SearchBot
- Used for
- Surfacing sites in ChatGPT search
- If you block it
- Your site is kept out of ChatGPT search answers
GPTBot
- Used for
- Training OpenAI's foundation models
- If you block it
- Your content should not be used for training. OpenAI does not say this affects search
ChatGPT-User
- Used for
- Visiting a page when a user asks
- If you block it
- Not used to decide what appears in search
PerplexityBot
- Used for
- Surfacing and linking sites in Perplexity results
- If you block it
- Your site may not be surfaced in Perplexity
Perplexity-User
- Used for
- Fetching a page a user asked about
- If you block it
- Perplexity says it generally ignores robots.txt, since a user requested it
OpenAI says each setting is independent, so you can allow OAI-SearchBot to appear in search while disallowing GPTBot for training. Perplexity also recommends allowing its published IP ranges. If you use a firewall or CDN bot filter, check it too, because robots.txt is not the only thing that can block a crawler.
What makes a page worth citing?
The best public evidence is a research paper from 2024, GEO: Generative Engine Optimization. In its benchmark, content changes such as adding citations, quotations from relevant sources and statistics raised a source's visibility by up to 40%, and which method worked best varied by domain. The authors report that keyword stuffing did little.
Treat that as a direction, not a formula. It was measured on the paper's own visibility metric, and these products have changed since. The practical reading is sensible anyway: a page with a specific fact and a named source gives an AI system something it can attribute to you.
- State the answer first. Put the direct answer at the start of the section, in plain language.
- Be specific. Prices, steps, limits and dates are more useful than general claims.
- Name your sources. Link to the original document, not to someone else's summary of it.
- Show who you are. A clear About page and consistent business details help a system identify your brand.
- Keep it current. Show the date and update pages when the facts change.
Do not fake statistics or mentions
If citations and statistics help, it is tempting to invent them. Do not. A wrong number attached to your brand does lasting damage, and Google says that inauthentic mentions are not as helpful as they might seem. Use your own real data, or link to someone else's.
How can you tell if it is working?
There is no ChatGPT equivalent of Search Console in the crawl and ranking sense, so measure what you can. Look for referral visits from chatgpt.com and perplexity.ai in your analytics, and ask a fixed list of buyer questions in each tool once a month. Record whether you are cited and which page they used. Answers vary between runs, so look at the trend, not a single result.
Be reachable, be specific, and name your sources. Anything beyond that is a test, not a known rule.
Read our guide to SEO, AEO and GEO for how this fits with Google, or see how AI Blogs Publishing can keep a site supplied with structured articles.
Sources
Questions, answered.
Does blocking GPTBot remove my site from ChatGPT search?
OpenAI's documentation treats them as separate. OAI-SearchBot controls whether a site appears in ChatGPT search, and GPTBot relates to model training. OpenAI says each setting is independent.
How long do robots.txt changes take to apply?
OpenAI and Perplexity both say changes can take about 24 hours to apply to their search systems.
Can an agency guarantee ChatGPT citations?
No. Neither OpenAI nor Perplexity publishes how sources are chosen, and answers change between runs. Treat any guarantee as a warning sign.
Do I need llms.txt to be cited by ChatGPT?
OpenAI's crawler documentation does not list it as a requirement. It is optional and low cost, but allowing OAI-SearchBot matters far more.
Related articles
llms.txt: What It Is and Whether You Need One
llms.txt is a Markdown map of your site for AI assistants. What goes in it, how it relates to robots.txt and sitemaps, and whether your site needs one.
What Is Generative Engine Optimization (GEO)?
GEO is how you become a source that ChatGPT, Perplexity and Gemini cite. What it is, how it differs from SEO and AEO, and the tactics that work today.
How to Appear in Google AI Overviews
To appear in Google AI Overviews, a page must be indexed and eligible for a snippet. What Google says helps, what it says you can skip, and how to measure it.