How to get cited by Perplexity, the 2026 playbook

Perplexity is the most citation-transparent of the major AI search engines. Every answer surfaces a numbered list of sources, every source is clickable, and the retrieval signals are observable in a way ChatGPT and Google AI Mode hide from the user. If you understand which signals matter to Perplexity, you can move the needle measurably inside 30 to 60 days.

This guide breaks down how Perplexity selects citations, the bot agents it sends, the retrieval backbone it uses, and the 8 steps that earn consistent placement.

How Perplexity picks citations

Perplexity uses its own retrieval index, branded as "sonar-online." The index is built by PerplexityBot crawling the open web and is augmented at query time by Perplexity-User, the agent that fetches specific URLs during a live conversation.

When a user asks a question, Perplexity routes through three stages:

1. Query decomposition. The question is broken into sub-queries that each retrieve a candidate document set.

2. Document ranking. The retrieved documents are re-ranked by topical relevance, source authority, recency, and structural fit (does the document contain a quotable answer span).

3. Answer synthesis with inline citation. The model generates an answer, attributing each factual claim to a numbered source from the re-ranked set.

The structural fit step is where most brands lose. Perplexity strongly prefers documents where a quotable 1 to 3 sentence answer span can be extracted under a clear header. Long documents that bury answers underperform short documents that lead with them.

Perplexity's bot user agents

Two agents matter.

Allowlist in robots.txt:

```

User-agent: PerplexityBot

Allow: /

User-agent: Perplexity-User

Allow: /

```

Perplexity publishes bot details at docs.perplexity.ai/guides/bots. Verify your CDN does not silently 403 either agent (Cloudflare's "Block AI Scrapers and Crawlers" toggle blocks both by default).

What Perplexity cites most

OnyxRank ran 3,800 Perplexity scans across diverse buyer-intent queries in April and May 2026. Citation distribution by source type:

Reddit dominance is even stronger on Perplexity than ChatGPT. Tinuiti's Q1 2026 study (tinuiti.com) measured Reddit at 24% of all Perplexity citations, exactly matching our independent finding. This is not a temporary bias; Perplexity has explicitly invested in conversational and forum-style sources because they contain the question-answer pairs the retrieval system was trained to find.

Notably, brand-owned content fares better on Perplexity (10%) than on ChatGPT (6%). Long-form blog posts that are well-structured do get cited; the citation rate is higher than on any other engine if you publish for the format.

The 8-step Perplexity action plan

1. **Allowlist both bots and verify with curl.** Run `curl -A "PerplexityBot" https://yourdomain.com` and confirm 200, not 403.

2. **Audit your top 20 pages for answer-span extractability.** Open each page, find the H2 that mirrors the implied buyer question, and check the first 1 to 3 sentences under it. If they do not contain a self-contained, quotable answer, rewrite.

3. **Add a sources block to every long-form post.** Perplexity favors documents that themselves cite primary sources. A "Sources" or "References" section at the bottom of long-form content increases authority signal.

4. **Earn Reddit citations.** 24% of Perplexity citations come from Reddit. Identify the 5 subreddits where your buyers ask questions. Build comment history over 90 days before referencing your brand. Quality and patience beat volume.

5. **Publish comparison content.** Pages structured as "X vs Y" or "best X for Y" get cited disproportionately because they directly answer the comparison questions Perplexity routinely receives. OnyxRank publishes 20 such comparison pages; we observe 3x the citation rate on those pages compared to general blog content.

6. **Get on niche forums.** Hacker News, Stack Exchange, Indie Hackers, and category-specific forums account for 6% of citations. A single well-received Hacker News comment can result in 50+ Perplexity citations over the following 60 days as the page gets re-indexed and re-cited.

7. **Submit to /llms.txt and /llms-full.txt.** Perplexity's crawler honors these files. A 200,000-character /llms-full.txt with your full blog corpus inlined increases the surface area Perplexity can sample for citations by an order of magnitude.

8. **Run weekly scans against 15 buyer prompts.** Track citation rate, citation quality (primary recommendation vs passing reference), and competitor share. Use https://onyxrank.com/tools/citation-checker for a free run.

Common mistakes

**Treating Perplexity as a small channel.** Perplexity has 50+ million monthly active users as of Q1 2026 (perplexity.ai/about). It is not a tier-3 engine; it is a meaningful share of AI-mediated buyer research, especially in B2B.

**Optimizing for snippet length but not snippet quality.** A 50-word answer that hedges and qualifies is worse than a 20-word answer that takes a position. Perplexity favors direct, attributable answers over committee-edited prose.

**Ignoring Reddit because "we cannot control it."** You cannot control Reddit, but you can earn the right to be referenced on it. Brands that have native presence on the 5 subreddits where their buyers gather will out-cite brands that rely entirely on owned content.

**Treating /llms-full.txt as a brochure.** This is not a marketing page. It is the LLM-ingestible knowledge base. Inline blog post bodies, comparison details, methodology. The biggest /llms-full.txt files in the AI SEO category exceed 1 megabyte for a reason.

How to verify you are working

Three measurable checks.

1. **Citation rate.** Run the same 15 prompts weekly. Track citation count and citation quality.

2. **Reddit referrer growth.** Watch the (Reddit) referrer in Google Analytics or your alternative analytics tool. Healthy growth correlates with healthy Perplexity citation share.

3. **PerplexityBot fetch frequency.** Grep access logs for the PerplexityBot user agent. A healthy site sees fetches at least 3 to 5 times per week across the most active pages.

What to ship next

Once the 8 steps are in motion, the highest leverage next step is a quarterly benchmark study or original data report. Perplexity cites primary-data documents at a much higher rate than secondary commentary. Publish one defensible data study per quarter and the citation share compounds.

Free citation audit at https://onyxrank.com/tools/citation-checker. Full strategy across all engines at https://onyxrank.com/blog/ai-citation-formula-geo-optimization-2026.