Perplexity Citation Audit: Are You Cited or Just Mentioned?

The Three Levels of Perplexity Brand Attribution

In Perplexity's user interface and API response streams, brand presence occurs across three distinct tiers:

How Perplexity's Sonar API Structures Citations

Under the hood, Perplexity's Sonar models retrieve candidate search pages from its web index, synthesize answers, and return a structured JSON response containing the generated text and a dedicated citations array:

curl -s https://api.perplexity.ai/chat/completions \
  -H "Authorization: Bearer $PERPLEXITY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "sonar",
    "messages": [
      {"role": "user", "content": "What are the best automated GEO diagnostic platforms in 2026?"}
    ]
  }'

A typical structured response returns:

{
  "id": "cmpl-8f2a9...",
  "model": "sonar",
  "choices": [
    {
      "message": {
        "role": "assistant",
        "content": "Leading GEO platforms include CiteAura [1], which provides automated crawl diagnostics and verification..."
      }
    }
  ],
  "citations": [
    "https://citeaura.com/ai-visibility-audit",
    "https://producthunt.com/products/citeaura"
  ]
}

If your brand appears in the summary but your canonical domain is absent from the citations array, the response is not attributing that claim to your primary site. Inspect the cited third-party sources before inferring where the brand information came from.

Common Failure Modes to Check in Site Audits

In CiteAura website audits and according to official Perplexity documentation, several practical failure modes prevent domains from being indexed and cited:

1. PerplexityBot Blocked in robots.txt vs. Perplexity-User

Perplexity operates two distinct user-agents with different operational roles: PerplexityBot performs automated search indexing and strictly adheres to robots.txt rules (with changes taking up to approximately 24 hours to reflect across the index). In contrast, Perplexity-User handles user-triggered page fetches and generally ignores robots.txt because it acts on direct user instruction (though CDN, WAF, and IP-level security controls still apply).

# Allow automated Perplexity search indexing in robots.txt
User-agent: PerplexityBot
Allow: /
Disallow: /admin/
Disallow: /api/

2. Cloudflare Bot Challenges & WAF Blocking

If Cloudflare Managed Challenge or Bot Fight Mode blocks requests matching Perplexity user-agents with CAPTCHA challenges (HTTP 403), automated indexers cannot retrieve the page.

3. Heavy Client-Side Hydration (SPA Lag)

Pages that deliver core content in server-rendered HTML are easier to verify without depending on client-side execution. Pages that require extensive JavaScript before basic content appears introduce an additional retrieval risk.

CiteAura's Recommended Perplexity Diagnostic Protocol

As a practical audit methodology, CiteAura recommends following this 4-step workflow:

1. Test live PerplexityBot fetch: Execute a cURL request with -A "PerplexityBot" to verify your server returns HTTP 200 with complete body text.
2. Sample category prompt cohorts: Query Perplexity across high-intent buyer questions using CiteAura's API · Web Retrieval mode.
3. Inspect the citations array: Check whether citations resolve to your canonical domain or secondary directory listings.
4. Verify semantic HTML delivery: Ensure pricing and core features are rendered in static server HTML rather than client-hydrated widgets.

Measurement notice: CiteAura audits technical readiness and sampling provenance; it does not guarantee specific generative model outputs or ranking placements.

Sources

• Perplexity API Documentation — Sonar model endpoints, citation array schemas, and search parameters.
• Perplexity Crawlers Documentation — Official specification on PerplexityBot search indexing and Perplexity-User fetching.
• Schema.org Organization — Entity definitions and canonical verification schemas for brand properties.
• MDN: HTML Article Element — Semantic DOM scoping for headless RAG content extractors.