To rank in Perplexity, let PerplexityBot, the bot that reads web pages for Perplexity, read your site. Then answer the buyer's question on one page. Get your brand named on the websites Perplexity already cites for your category. Keep your facts the same everywhere, and check the answers every week.
Perplexity is an AI answer engine. It searches the web when you ask a question and writes one answer with links to its sources (Perplexity help centre, checked 7 Oct 2026). Ranking in Perplexity means being one of those sources, or being a brand the answer names.
Perplexity documents how its bots behave. We found no guide from Perplexity for site owners on how it chooses which pages to cite. So statements from Perplexity below link to its own pages, and Depra AI's measurements are labelled as observation.
How does Perplexity answer a question?
Perplexity searches the web at the moment you ask, then writes an answer from what it finds. Its help centre says it searches the internet in real time, and that every answer includes citations linking to the original sources (Perplexity help centre, checked 7 Oct 2026).
Perplexity also sells its search to developers through an API, an interface that software uses to call a service. Its documentation says the Search API returns real-time, ranked web results (Perplexity Search API documentation, checked 7 Oct 2026).
Its API FAQ adds that the API uses the same search system as the app with differences in configuration, so their outputs may differ (Perplexity API FAQ, checked 7 Oct 2026). That matters when a tool tracks Perplexity through the API.
What does Perplexity say about PerplexityBot, Perplexity-User and robots.txt?
Perplexity documents two bots that visit websites. PerplexityBot is its search crawler, and it follows robots.txt. Perplexity-User fetches a page when a person's question needs it, and Perplexity says it generally ignores robots.txt (Perplexity crawler documentation, checked 7 Oct 2026).
A crawler is a bot that visits web pages and reads them. The robots.txt file is a plain text file at yourdomain.com/robots.txt that tells bots which pages they may fetch.
- Perplexity recommends allowing PerplexityBot. Its crawler page says that to appear in its search results you should allow PerplexityBot in robots.txt. It also asks you to permit requests from its published IP ranges, the network addresses its bot uses (Perplexity crawler documentation, checked 7 Oct 2026).
- A block does not hide you fully. Perplexity's help centre says it does not index the text of a page that robots.txt blocks. It may still index the domain, headline and a brief factual summary (Perplexity help centre on robots.txt, checked 7 Oct 2026).
| Bot | What Perplexity says it does | robots.txt | AI training |
|---|---|---|---|
| PerplexityBot | Surfaces and links websites in search results on Perplexity | Follows it. The text of a blocked page is not indexed, but the domain, headline and a brief factual summary may be | Not used to crawl content for AI foundation models |
| Perplexity-User | Visits a page when a user's question needs it, and may link to that page in the answer | Generally ignores it, because a user requested the fetch | Not used to collect content for training AI foundation models |
What kinds of pages does Perplexity cite?
Perplexity gives site owners no checklist that we could find. Its help centre describes its sources only in general terms, as authoritative sources such as articles, websites and journals (Perplexity help centre, checked 7 Oct 2026). An engineering article on its site says its search API ranks in several stages. It also says the API treats sections and spans of a document, meaning stretches of text, as units in their own right (Perplexity research article, checked 7 Oct 2026). What follows is observation, not a rule from Perplexity.
Depra AI's own study gives one measured view. We asked 10 buying questions about skincare and fashion on ChatGPT, Gemini and Perplexity from India on 14 Aug 2026. Each ran 8 times in English and 8 times in Hinglish, which is Hindi and English typed in English letters. That made 480 answers, with Perplexity measured through its API. Full method: the Hinglish vs English AI shopping study.
- Perplexity cited sources every time. It attached sources to all 80 English answers and all 80 Hinglish answers.
- Reruns were the most stable. On average, two Perplexity answers to the same question in the same language shared 68.3% of the brands they named. That was the highest of the three engines.
- Language moved the list. An English and a Hinglish answer to the same question shared 44.9% of their brands. That 23.4 point drop was the largest of the three.
- Video, forums and marketplaces led the citations. Across all three engines, the 480 answers carried 4,640 citations. youtube.com had 534, reddit.com 191 and nykaa.com 176. Small Indian review blogs were cited more often than most mainstream publishers.
How to read these numbers
They cover two categories on one day, and the citation counts combine three engines. For your own category, open the sources under ten Perplexity answers and sort them by type.
How do you rank in Perplexity? A 5-step plan
Work in this order: access, your own page, other sites, consistent facts, then measurement. Step 1 follows Perplexity's own documentation. Steps 2 to 5 are our recommendation. How to get ChatGPT to recommend your brand covers the same four steps in more depth. No step guarantees a citation, because Perplexity decides what to cite.
Step 1: Let the crawler in
Open yourdomain.com/robots.txt. Look for a Disallow rule under User-agent: PerplexityBot, or under User-agent: *, the group that covers every bot. If a rule blocks the bot, remove it. Or add a group with two lines, User-agent: PerplexityBot and Allow: /.
Then check your firewall and your CDN, the service that delivers your pages, such as Cloudflare. A bot filter there can refuse PerplexityBot with a 403 error, the "access denied" reply, while robots.txt allows it. To check, search your server logs for PerplexityBot and look at the status code on each request. Perplexity's crawler page has setup notes for Cloudflare and AWS firewalls (Perplexity crawler documentation, checked 7 Oct 2026).
That page does not say whether PerplexityBot runs JavaScript, the code that loads content after a page arrives. So put prices, sizes and ingredients in the page's own text, where a bot can read them without running a script. Depra AI's AI crawler checker shows whether your robots.txt allows each of 41 AI crawlers, PerplexityBot included.
Step 2: Answer the question on one page
Pick one buying question, such as "best sunscreen for oily skin under ₹500", and give it one page. Put the question in the heading and the answer directly under it. Then give the price, the buyer it suits, the buyer it does not suit and how it compares.
Write the answer as text, with numbers a reader can check. Perplexity says its search API treats sections of a document as units (Perplexity research article, checked 7 Oct 2026). So let each section answer one question fully.
Step 3: Earn mentions on the sites Perplexity already cites
Ask Perplexity your top 10 buying questions and open every source under each answer. The pages that name your competitors and not you are your target list. Depra AI's citation gap analysis lists the sites cited in stored answers that name a competitor while your brand is missing. The pricing page lists it on Growth and above.
- List articles: write to the author with your price and ingredients, and offer a sample.
- Forum threads: answer open questions from a named brand account, within the rules of each forum.
- Marketplaces and directories: create or claim your listing so the page exists for an engine to find and cite.
- Video reviews: offer the product to reviewers who already cover your category.
Step 4: Keep your facts the same everywhere
Use one product name, one price, one pack size and one description on your site, your marketplace listings and your profiles. An answer built from pages that disagree can pass on the wrong one. So when a price changes, update every listing together.
Step 5: Measure every week
Count the share of answers that name you, week by week. The next section gives the method. One answer proves little. In our study, two Perplexity reruns of the same question shared 68.3% of their brands, so a single answer can leave you out by chance.
How to measure AI visibility explains how many answers you need.
How do you check Perplexity by hand?
You can sample Perplexity yourself before you pay for a tool. The spreadsheet method for tracking brand mentions in AI answers gives you the columns.
- 1. Write 10 buying questions. Leave brand names out, and use the words your buyers use.
- 2. Ask each one in a new thread. Use perplexity.ai and note which model or mode is selected.
- 3. Ask each question 3 times. Reruns return different lists.
- 4. Log five things. Named or not, your position among the brands, the tone, the sentence that names you, and every source link.
- 5. Repeat weekly. Keep the wording fixed and compare the share of answers that name you.
How does Depra AI track Perplexity?
Depra AI checks Perplexity once a week on every plan. It asks your tracked questions through Perplexity's API, with live web search set to the country you track. It stores every answer with the brands named, your position, the tone of the mention and the websites cited.
This is an API capture. Depra AI does not record the Perplexity app, where the model and settings can differ. An answer you see in the app may not match the stored one. Location is set by country, in any of 40 markets, with no city targeting.
ChatGPT, Gemini and Google AI Overviews are checked daily in the same project. Questions can be tracked in English and Hinglish. The Perplexity visibility tracker page explains the weekly schedule and what the report shows.
Depra AI does not write your pages, change your site or promise a citation. Plans start at ₹1,999 a month plus GST, and Perplexity is included in every plan. See Depra AI pricing.
Does the same work help on ChatGPT, Gemini and Google AI Overviews?
Mostly yes. Steps 2 to 5 are the same on every engine: clear answers, mentions on cited sites, consistent facts and repeated measurement. Step 1 changes, because each engine has its own crawler or control.
- ChatGPT: how to get ChatGPT to recommend your brand covers its search crawler. If you are missing there, start with why ChatGPT does not mention your brand.
- Gemini: how to rank in Gemini covers the Google Search index and the Google-Extended rule in robots.txt.
- Google AI Overviews: how to rank in Google AI Overviews covers Google's own rules for its AI summaries.
- All engines: the AI SEO guide puts the steps in one plan, and AI visibility explains the number you are trying to move. Brand monitoring in AI answers covers what to log in each answer.