How to get cited by Perplexity and ChatGPT search
Let the right crawlers in, make pages readable without JavaScript, answer questions plainly and earn mentions elsewhere. The documented steps.
- perplexity
- chatgpt
- citations
To be cited by Perplexity and ChatGPT search, three things have to be true: their search crawlers can fetch your pages, the pages are readable as plain HTML, and a page answers the question more clearly than the alternatives. OpenAI and Perplexity document the first one precisely. The other two rest on engineering basics and on limited research, and nobody can guarantee a citation.
Below: the steps that are documented, the ones that are plausible, and how to check each. For the broader picture, see how assistants decide which brands to recommend.
Which crawlers do I need to allow?
OpenAI and Perplexity each run separate bots for search, user-initiated fetches and (for OpenAI) training:
| Bot | Company | What it is for | robots.txt |
|---|---|---|---|
| OAI-SearchBot | OpenAI | Surfacing sites in ChatGPT's search features | Respected; OpenAI recommends allowing it |
| ChatGPT-User | OpenAI | Fetches triggered by a user action | Rules "may not apply" per OpenAI |
| GPTBot | OpenAI | May be used to train generative AI models | Respected; block it to opt out of training |
| PerplexityBot | Perplexity | Surfacing and linking sites in Perplexity results | Respected; Perplexity recommends allowing it |
| Perplexity-User | Perplexity | A user asks a question and the page is fetched | "Generally ignores" robots.txt |
Sources: OpenAI bots and Perplexity bots. OpenAI states that each setting is independent, so you can allow OAI-SearchBot while disallowing GPTBot, and that sites that opt out of OAI-SearchBot will not appear in ChatGPT search answers, although they can still appear as navigational links. Perplexity says PerplexityBot is not used to crawl content for AI foundation models.
A robots.txt that allows both search bots and declines OpenAI training:
User-agent: OAI-SearchBot
Allow: /
User-agent: PerplexityBot
Allow: /
User-agent: GPTBot
Disallow: /
Whether to block training is your call; it does not affect search inclusion, according to OpenAI. Both vendors say robots.txt changes can take effect within about 24 hours. Our free AI crawler checker reads your robots.txt and tells you which bots are blocked.
Could something else be blocking the bots?
Yes, and this is a common surprise. robots.txt is only one layer. A CDN or firewall can block bots by user agent or IP before they ever read your robots.txt. Cloudflare announced in July 2025 that it was changing its default to block AI crawlers unless they pay creators (Cloudflare). Check your CDN's bot settings, and if you use a web application firewall, Perplexity recommends matching both the user agent string and its published IP ranges, and OpenAI recommends allowing requests from its published ranges. Perplexity publishes JSON lists for PerplexityBot and Perplexity-User, and OpenAI documents ranges for its bots.
How to check: look in your server or CDN logs for these user agents with a 200 status. If you see 403 or 429, fix that first. Nothing else on this list matters until it works.
Does my page need to work without JavaScript?
Safest to assume yes. A late-2024 study by Vercel and MERJ of AI crawler traffic found that GPTBot, ClaudeBot and PerplexityBot did not render JavaScript: they sometimes fetched JavaScript files but did not execute them, so content that only appears after client-side rendering was invisible to them (Vercel). That is a snapshot from December 2024 and crawlers change, so verify for the bots you care about. But the safe engineering answer does not depend on it: serve your main content in the initial HTML.
Test it: curl -s https://yourdomain.com/your-page | grep -i "a phrase from your main text". If the phrase is missing, a crawler that does not run scripts will not see it either. Google also recommends server-side or pre-rendering as still worthwhile because not every bot can run JavaScript (Google).
What makes a page worth citing?
No vendor publishes a rule, but there is some evidence on writing style. The GEO research paper (Aggarwal et al., KDD 2024) tested content changes against generative engines including Perplexity.ai. Adding citations to reliable sources, relevant quotations and statistics improved visibility by roughly 30 to 40 percent on one metric and 15 to 30 percent on another in the main benchmark, while keyword stuffing gave little to no improvement (arXiv). The Perplexity.ai test was a separate, smaller experiment. It is an academic benchmark, not a promise for your pages.
Practical rules that follow without overclaiming:
- Lead with the answer. The first two or three sentences under the H1 should answer the question in plain words.
- Use question-style headings that match how buyers phrase things.
- Support claims with a named source, a date and a number, linked.
- Say who you are for and who you are not for, plus price and limits. Vague pages are hard to quote.
- Add a real FAQ with the questions sales actually hears.
- Date and update pages. A visible last-updated date helps readers; whether engines weigh it is not documented, so do not rely on it.
Do I need to be on other sites too?
Probably, for "best tool" questions. In DerivateX's study published 31 May 2026 (40 B2B SaaS categories, ChatGPT with web search, each buyer prompt run ten times), ChatGPT cited a recommended tool's own website 11.6 percent of the time and third parties the rest (another section of the report gives 12.6 percent for the same measure), with G2 and Capterra receiving no citations across 233 recommendations (DerivateX). The sample is small, it covers one engine and one query type, and the publisher sells GEO services. Still, it fits what you see if you read the sources of your own category: roundups, comparisons and community discussions.
What you can do without gaming anything: get listed or reviewed honestly where your buyers read, publish a fair comparison page on your own site, and answer questions in communities where you are a real participant. Do not buy fake reviews and do not mass-post links; both hurt people and tend to backfire.
How do I know whether it worked?
Two separate checks:
- Visits: ChatGPT adds the parameter
utm_source=chatgpt.comto links in its search results, according to OpenAI's publishers FAQ, so analytics tools can attribute those clicks. - Mentions and citations: clicks undercount. An assistant can name you without a click. You need to ask the questions repeatedly and record whether you appear.
What do we measure, and how can you check?
Nodpick is built to ask your buyer questions in ChatGPT and Perplexity (the free check also covers Gemini and Claude; paid plans add Google AI Overviews and AI Mode) and record mention, position and cited domains. The citation list will show which pages Perplexity or ChatGPT used instead of yours. We run them through the vendors' programming interfaces, so results can differ from what a person sees in the apps; our method page says so plainly. Results are estimates, not guarantees.
Nodpick is not open yet. Its free check will show which of your buyer questions name you and which domains get cited instead; join the waitlist and we email you once, the day it opens.
FAQ
Does blocking GPTBot remove me from ChatGPT? According to OpenAI, no. GPTBot relates to training; OAI-SearchBot relates to search results.
Can a blocked page still show up? OpenAI says sites that opt out of OAI-SearchBot will not be shown in ChatGPT search answers but can still appear as navigational links.
How long until changes show? OpenAI and Perplexity both say robots.txt changes can take about 24 hours to be reflected. Being named more often takes longer and depends on third-party pages; see the GEO guide for small teams.
Do I need an llms.txt file? Evidence of benefit is missing; see our llms.txt guide.
Can I pay Perplexity or OpenAI for a citation? We found no such product in the documentation we cite. Be wary of anyone who sells one.