Quick answer: To get cited by ChatGPT and Perplexity, make your page easy for their crawlers to fetch and easy for the model to quote. That means an answer-first page in clean server-rendered HTML, self-contained quotable sentences, clear structure, and enough topical authority that the engine trusts you. Both tools search the live web, read a few pages, and cite the ones that answer the query most directly. Your job is to be that page.
When ChatGPT or Perplexity answers a question with links, those links are not decoration. The engine ran a search, fetched real pages, and built its answer from what it read. Being one of the cited sources puts your brand in front of the user at the exact moment they are getting an answer. That is valuable, and unlike classic SEO, only a handful of sources make the cut per query.
So how do you become one of them? Start with how the fetch actually works.
How ChatGPT and Perplexity fetch sources
Both tools follow a broadly similar loop, even though the details differ and are proprietary.
The user asks a question. The engine turns that into one or more search queries, often several reformulations for a single question. It retrieves a candidate set of URLs. Then it fetches those pages, reads the content, and synthesizes an answer, attaching citations to the sources it relied on most.
The important consequence: there are two gates. First your page has to be retrievable, meaning it surfaces for the search the engine runs. Then it has to be fetchable and readable, meaning the crawler can pull your content and the model can extract a clean answer from it. Plenty of pages clear the first gate and die at the second because their content is locked behind JavaScript or buried under filler.
We walk through this fetch-and-read loop in detail in what happens when ChatGPT searches the web. The gist for citations: the engine can only cite what it can read, and it prefers to cite what it can read cleanly.
What makes a page citation-friendly
Think about the model's job at the synthesis step. It has several pages open and needs to write one answer with citations. It will lean on the source that gives it a clean, correct, quotable claim with the least effort. Optimize for that.
| Property | Why it earns a citation |
|---|---|
| Answer-first | The model finds a usable claim immediately |
| Quotable sentences | Statements survive being lifted out of context |
| Clean structure | Sections map to sub-questions the model is answering |
| Verifiable facts | Specific, sourced claims read as trustworthy |
| Crawler access | The content is actually in the fetched HTML |
| Freshness | Current pages win time-sensitive queries |
Answer first
Lead with the answer in plain language. If your page opens with a story or a definition of terms nobody asked about, the model reaches for a source that gets to the point. The quick-answer block at the top of this post is the pattern.
Write sentences that survive being quoted
A citable sentence stands alone. "Perplexity retrieves several sources and synthesizes one answer" can be lifted straight into an AI response. "As mentioned, it does this too" cannot. Write declarative, self-contained sentences for your key claims.
Make the facts checkable
Both engines lean toward sources that look reliable. Name your sources, cite your data, and be specific about what, who, and when. Specificity reads as authority. Vague hand-waving reads as filler and gets skipped.
Make sure the crawler can read you
This is the one people miss. If your answer only appears after client-side JavaScript executes, some crawlers never see it, and a page the model cannot read is a page it cannot cite. Serve your core content as real HTML.
You can check what a crawler receives by fetching the raw page yourself:
curl https://link.sc/v1/fetch \
-H "Authorization: Bearer lsc_your_key" \
-H "Content-Type: application/json" \
-d '{
"url": "https://yoursite.com/your-page",
"format": "markdown"
}'
If the returned markdown is missing your answer, that is what the AI crawler sees too. Clean, parseable content is the whole reason tools like link.sc exist, and the same property that makes a page easy for a fetch API makes it easy for ChatGPT to cite.
How to check if you are being cited
You cannot manage what you do not measure, so build a simple citation-tracking habit.
- Ask directly. Put your target questions to ChatGPT (with search on) and Perplexity, and record whether you appear in the citations.
- Track over time. Re-run the same questions on a schedule. Citations shift, so a single check is a snapshot, not a trend.
- Watch referrals. AI engines send a small but growing stream of referral traffic. Filter your analytics for it.
- Study the winners. For every query where a competitor is cited and you are not, fetch their page and compare structure, clarity, and freshness against yours.
That last step scales well with an API. You can pull the cited pages as clean markdown and diff them against your own content programmatically, instead of squinting at a dozen browser tabs.
Tactics that move the needle
Here is the short list, in rough priority order:
- Rewrite your top pages answer-first. Biggest lever, lowest cost.
- Fix crawlability. Confirm the answer is in the raw HTML on every important page.
- Break pages into clean, self-contained sections with question-shaped headings.
- Add specificity. Replace vague claims with sourced, dated facts.
- Refresh regularly. Re-date and update pages that cover changing topics.
- Build authority by covering your topic deeply and earning references elsewhere.
Notice none of these are tricks. They are the same moves that make content genuinely good, re-weighted for machines that read before they answer.
An honest caveat
The selection logic inside ChatGPT and Perplexity is proprietary and changes often. Citation results vary between sessions, models, and even identical prompts. Anyone promising guaranteed citations is guessing. What we can say with confidence is that clear, quotable, crawlable, current, authoritative content is what both engines reward today, and none of those properties will hurt you if the mechanics change tomorrow.
Write for the reader who wants the answer now. Increasingly, the reader is a model, and the model cites the page that helps it most.
Check whether ChatGPT and Perplexity can even read your key pages. Fetch any URL to clean markdown with link.sc and see what the crawler sees.