Perplexity Spaces: optimization for the multi-source Perplexity research surface

A comprehensive installation and audit reference for winning citation and brand visibility on Perplexity's surface family. Perplexity Search (the original answer engine, launched December 7, 2022 per…

Earning Citations and Visibility on Perplexity's Discovery Surfaces: Standard Search, Pages, Spaces, and the Comet Agentic Browser

A comprehensive installation and audit reference for winning citation and brand visibility on Perplexity's surface family. Perplexity Search (the original answer engine, launched December 7, 2022 per Wikipedia and Britannica Money coverage of the founding by Aravind Srinivas, Denis Yarats, Johnny Ho, and Andy Konwinski), Perplexity Pages (long form research outputs published as standalone URLs, launched May 30, 2024 per the Perplexity Pages launch blog), Perplexity Spaces (saved research bundles with custom system prompts and uploaded files, launched October 2024 per the Internal Knowledge Search and Spaces announcement), and Perplexity Comet (the Chromium based agent browser, launched July 9, 2025 per the Comet launch blog) share a bot family, a retrieval architecture, and a citation transparency philosophy, but each surface produces distinct citation behavior. This framework specifies the structural, schema, freshness, entity, and access signals that drive citation on those surfaces, the bot family and how each member behaves, the audit rubric, and the maintenance cadence. Dual purpose: installation manual and audit document.

Cross stack implementation note: code samples in this framework are written in plain HTML for clarity. For React, Vue, Svelte, Next.js, Nuxt, SvelteKit, Astro, Hugo, 11ty, Remix, WordPress, Shopify, and Webflow equivalents see framework-cross-stack-implementation.md. For pure client rendered SPAs (no SSR/SSG) see framework-react.md. For Tailwind specific concerns see framework-tailwind.md.

1. Document Purpose

1.1 What This Document Is

This is the canonical operational reference for citation on Perplexity's surface family. Perplexity grew from approximately 2 million monthly active users in March 2023 to over 45 million monthly active users by early 2026 per the Panto AI 2026 Perplexity statistics aggregation. Annualized recurring revenue hit approximately 500 million USD by April 2026 per the Sacra 2026 tracker, up roughly 335 percent year over year. By mid 2026 the platform processes 1.2 to 1.5 billion search queries per month per Wytlabs's 2026 roundup.

The Perplexity surface family is structurally distinct from the OpenAI family and Google AI Overviews. Different bot family. Different retrieval philosophy (RAG with prominent inline citations). Different source preference (heavy Reddit and primary source weighting). Different citation rate (8.79 per response per BrightEdge 2025 to 2026 Perplexity guide, against 3 to 7 for ChatGPT Search). Different surface architecture: standard search produces transient answers; Pages produces durable Google indexable URLs; Spaces produces persistent research workspaces; Comet produces agent driven task execution.

Per the Whitehat SEO 2026 AI engines comparison and the Profound 2025 platform citation patterns analysis, only approximately 11 percent of domains are cited by both ChatGPT and Perplexity. This document covers the Perplexity specific signals.

For broader multi engine AI citation work see framework-aicitations.md. For ChatGPT Search see framework-searchgpt.md. For Google AI Overviews see framework-aioverviews.md. For the substrate doctrine see framework-contentfirst.md.

1.2 Three Operating Modes

Mode A, Install Mode. Build Perplexity ready infrastructure on a new or existing site. Follow Sections 2 through 14 in order.

Mode B, Audit Mode. Evaluate an existing site for current Perplexity citation state. Skip to Section 13.

Mode C, Hybrid Mode. Audit first, then install for failing items.

1.3 How Claude Code CLI Should Consume This Document

1. Read Section 2 and collect client variables.
2. Run Section 13 audit on representative pages.
3. Apply Section 5 ranking signals to priority pages.
4. Configure Section 4 robots.txt and Section 10 llms.txt.
5. Apply Section 7 Pages playbook and Section 8 Spaces playbook where engagement scope permits.
6. Apply Section 9 Comet posture on commercial pages.
7. Establish Section 13.4 first 90 days tracking before measuring lift.
8. Apply Section 14 maintenance cadence and generate the Section 14 report.

1.4 Conflict Resolution Rules

1.5 Required Tools

• curl with custom user agent for substrate verification
• Server log access to identify PerplexityBot and Perplexity-User traffic
• An AI citation tracker: Otterly (lowest cost for Perplexity specific tracking per the Otterly 2026 25 AI SEO tools roundup), Profound, AthenaHQ, or equivalent
• Manual sampling discipline: weekly priority query sampling in Perplexity Search
• Google Search Console access (Perplexity often draws from Google indexed pages per the Skale 2026 Perplexity GEO guide)

1.6 Relationship to Neighboring Frameworks

This framework covers Perplexity Search, Pages, Spaces, and Comet. Broader AI citation lives in framework-aicitations.md. The OpenAI surface family lives in framework-searchgpt.md. Google AI Overview lives in framework-aioverviews.md. The substrate prerequisite lives in framework-contentfirst.md. The four pillars architecture lives in SEO-Search-Appearance.md. The agentic browsing playbook lives in framework-agenticaisearch.md.

2. Client Variables Intake

# PERPLEXITY FRAMEWORK CLIENT VARIABLES

# --- Business and Site Identity (REQUIRED) ---
business_name: ""
primary_domain: ""
business_industry: ""
ymyl_classification: ""                       # "full_ymyl" | "partial_ymyl" | "lite_ymyl" | "non_ymyl"

# --- Current Perplexity Citation State (REQUIRED) ---
perplexity_citation_status: ""                # "regularly_cited" | "occasionally_cited" | "rarely_cited" | "never_cited" | "unknown"
queries_currently_cited_in_perplexity: 0
priority_queries: []                          # 10 to 25 queries the site targets
avg_perplexity_citations_per_priority_query: 0 # benchmark approximately 8.79 per BrightEdge 2025
perplexity_pages_authored_by_client: 0
perplexity_pages_citing_client_site: 0

# --- Bot Access Status (REQUIRED) ---
perplexitybot_allowed: false                  # block this and indexing stalls
perplexity_user_allowed: false                # block this and real time citation fails
verified_via_server_logs: []
last_robots_txt_audit_date: ""

# --- llms.txt Status (RECOMMENDED) ---
llms_txt_exists: false
llms_full_txt_exists: false
llms_txt_last_updated: ""

# --- Content First Baseline (REQUIRED, see framework-contentfirst.md) ---
contentfirst_score: 0                         # out of 30
substrate_in_first_byte: false
schema_in_first_byte: false
js_required_for_primary_content: false

# --- Reading Mode (REQUIRED) ---
curl_perplexitybot_useragent_returns_full_html: false
heading_hierarchy_clean: false
faq_uses_details_summary: false
tables_for_comparison_data: false
ordered_lists_for_sequential: false

# --- Perplexity Source Posture (REQUIRED) ---
primary_source_links_on_priority_pages: 0
academic_or_gov_citations_count: 0
dateModified_visible_and_in_schema: false
content_refresh_cadence: ""
percent_priority_pages_updated_last_30_days: 0
percent_priority_pages_updated_last_90_days: 0

# --- Reddit and Forum (RECOMMENDED) ---
brand_presence_on_reddit: false
brand_mentions_in_reddit_threads: 0
youtube_transcripts_mentioning_brand: 0

# --- Schema and Entity (REQUIRED, see framework-schema.md) ---
core_graph_schema_server_rendered: false
article_or_blogposting_schema_with_dateModified: false
faqpage_schema_mirroring_visible_details: false
person_schema_for_authors_with_sameAs: false
organization_sameAs_includes_wikidata_qid: false

# --- Tracking (REQUIRED) ---
perplexity_citation_tracker: ""               # "Otterly" | "Profound" | "AthenaHQ" | "manual_only" | "none"
manual_sampling_cadence: ""
server_log_perplexity_bot_monitoring: false

Citation work cannot start until contentfirst_score is at least 22, PerplexityBot and Perplexity-User are allowed, and a baseline citation sample exists.

3. What Perplexity Spaces, Pages, and Comet Are

3.1 The Four Perplexity Surfaces

Perplexity operates a family of answer and research surfaces sharing one retrieval substrate.

Perplexity Search. The original answer engine, launched December 7, 2022 per Wikipedia and Britannica Money. The user types a question, Perplexity searches the open web, synthesizes an answer, and surfaces an inline citation set (5 to 15 sources; average 8.79 per BrightEdge 2025 to 2026). Dominant surface by query volume.

Perplexity Pages. Launched May 30, 2024 per the launch announcement and VentureBeat May 2024 coverage. A long form publishing surface where a query is transformed into a structured article with headings, media, and inline citations. Pages publish at perplexity.ai/page/[slug]-[id] and become Google indexable per the Search Engine Journal December 2024 analysis; a Page can outrank the underlying source.

Perplexity Spaces. Launched October to November 2024 per the Internal Knowledge Search and Spaces blog and Gilbane October 2024 coverage. A persistent research workspace with a custom system prompt, uploaded files (PDF, Word, Excel, PowerPoint, CSV; 50 for Pro, 500 for Enterprise), and shared access with viewer or contributor permissions. Threads accumulate context.

Perplexity Comet. Launched July 9, 2025 per the launch blog and IBM Think coverage. Desktop only for Max tier (200 USD per month) initially; Android November 2025; iOS March 2026 per the Comet Wikipedia entry. Chromium based browser with agent built into the chrome. Approximately 3 million MAU by early 2026 per SEO Sandwich, DAU up approximately 320 percent YoY.

3.2 Why These Are A Distinct Optimization Target

Three studies establish the divergence:

• Cross surface overlap is low. Per the Whitehat SEO 2026 AI engines citation comparison and the Profound August 2024 to June 2025 platform citation patterns analysis, only approximately 11 percent of domains are cited by both ChatGPT and Perplexity. Per the Discovered Labs 2025 to 2026 platform comparison, ChatGPT and Perplexity reach semantically similar conclusions in 86 percent of comparable prompts despite citing different sources in 89 percent of cases.
• Source preference diverges sharply. Per Profound's August 2024 to June 2025 analysis and the Otterly 1 million plus citation analysis January to February 2026, Perplexity cites Reddit at approximately 6.6 to 16.9 percent depending on sample window. Per the Tech Edge AI 2026 citation index, Reddit reached approximately 24 percent of Perplexity citations by January 2026 for queries where community discourse exists. Outside Reddit, Perplexity weights primary sources (NIH, PubMed, arXiv, peer reviewed journals, government data) higher than competitors per the Authority Tech 2026 Reddit strategy analysis.
• Citation rate diverges. Per BrightEdge's 2025 to 2026 Perplexity guide, the average Perplexity response cites approximately 8.79 sources. Per the Otterly 2026 Perplexity SEO guide, Perplexity is the most measurable AI citation channel because citations appear on every answer.

3.3 The 2026 State Of Perplexity

Five numbers define the surface for planning purposes:

1. Approximately 45 million MAU by early 2026, up from approximately 2 million in March 2023 (Panto AI 2026 statistics).
2. Approximately 4.4 to 6.6 million DAU in June 2025 (Panto AI 2026).
3. Approximately 1.2 to 1.5 billion queries per month by mid 2026 (Wytlabs 2026).
4. Approximately 500 million USD ARR by April 2026, up 335 percent YoY (Sacra 2026).
5. Approximately 3 million Comet MAU by early 2026, DAU up 320 percent YoY (SEO Sandwich 2026).

3.4 Four Pillars Context

Perplexity citation work sits in the GEO pillar of the four pillars architecture: SEO (classic ten blue links), AEO (Answer Engine Optimization), AIO (Google AI Overview, see framework-aioverviews.md), GEO (this framework and framework-aicitations.md and framework-searchgpt.md). A page can win SEO and fail Perplexity, or win Perplexity and fail SEO.

4. The Perplexity Bot Family and Crawl Behavior

Perplexity's bot family has two declared members as of Q2 2026, each controlled independently in robots.txt per Perplexity's crawler documentation. Treating them as one signal is the second most common Perplexity optimization mistake (after ignoring primary source attribution; see Section 12).

4.1 PerplexityBot (Index Crawler)

Builds and refreshes the index behind Perplexity Search answers per the official crawler documentation. Does not crawl content to train AI foundation models per the same documentation; analogous to OAI-SearchBot.

User agent: Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; PerplexityBot/1.0; +https://perplexity.ai/perplexitybot). IP ranges at https://www.perplexity.com/perplexitybot.json.

Blocking removes the site from Perplexity's persistent index. Crawl frequency varies by site popularity, content freshness, and query relevance per the documentation; no fixed schedule.

4.2 Perplexity-User (Live User Fetch)

Fetches pages in real time when a user submits a query requiring current data per the crawler documentation. Visits during the user's session.

User agent: Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; Perplexity-User/1.0; +https://perplexity.ai/perplexity-user). IP ranges at https://www.perplexity.com/perplexity-user.json.

Blocking prevents live re fetch verification. Allow both for full visibility.

4.3 Bot Family Summary

Default posture: allow both. Block only with a deliberate business decision to opt out of Perplexity visibility (see Section 4.6).

4.4 Recommended Robots.txt

# robots.txt

User-agent: PerplexityBot
Allow: /

User-agent: Perplexity-User
Allow: /

# All other rules
User-agent: *
Allow: /
Sitemap: https://example.com/sitemap.xml

The critical lines are the two Allow: / directives above. Without them, Perplexity Search citation is structurally impossible for the affected paths.

4.5 Verifying Bot Access Via Server Logs

Robots.txt is one signal; actual fetches in logs are the ground truth. Confirm at least monthly per Section 14.

grep -E "PerplexityBot|Perplexity-User" /var/log/nginx/access.log | \
  awk '{print $1, $7, $9, $12}' | sort | uniq -c | sort -rn | head -50

Status 200 means the bot read the page. 4xx or 5xx indicates a problem. The Section 14.4 script automates monthly reporting. For high authority sites, verify IPs match Perplexity's published ranges; reverse DNS is also supported.

4.6 The Undeclared Crawler Question

In August 2025, industry crawler reports documented behavior where Perplexity retrieved pages from sites that had blocked both declared bots, using rotating IPs and a generic Chrome on macOS user agent per the August 2025 industry analysis covered by Daring Fireball, Hacker News, Slashdot, Engadget, and Intelligent CISO. Observed across tens of thousands of domains.

This framework treats the declared bots as the contract. Sites wanting visibility have no action item; allowing the declared bots is sufficient. Sites wanting to block should set robots.txt directives, monitor logs (Section 14.4), and accept that detection of undeclared crawling requires WAF heuristics beyond this framework's scope.

4.7 Crawl Delay and Patterns

Perplexity bots respect Crawl-delay. Most sites need none; Crawl-delay: 5 is reasonable for constrained servers; above 30 seconds risks freshness penalty.

Per the Datadome 2025 profile and CrawlerCheck 2025 entries: initial discovery to first fetch is hours to days for sites with external links, weeks for new domains; re fetch frequency is every 7 to 14 days for evergreen, daily for actively updated content. New engagements should not expect visibility from day one; index freshness lags ChatGPT Search by approximately two to six weeks per the Skale 2026 GEO guide.

5. Perplexity Ranking Signals

Ordered by leverage (high to low) for Perplexity citation specifically. Each tagged as Shared with classic SEO or Perplexity specific.

5.1 Server Rendered Substrate (HIGHEST leverage, Perplexity specific)

None of Perplexity's declared bots execute JavaScript per the Datadome 2025 profile, the Known Agents reference, and the Hall AI 2025 summary. The first byte server response is the entire surface available for parsing. This is the single highest leverage signal; a site that fails the substrate test is invisible regardless of every other optimization. See framework-contentfirst.md and Section 5 of framework-searchgpt.md.

Triage test:

curl -A "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; PerplexityBot/1.0; +https://perplexity.ai/perplexitybot)" \
  -s https://example.com/priority-page/ | less

Look for: H1, lede paragraph, all H2s, FAQ content as actual text, comparison data in tables, procedural content in ordered lists, schema JSON-LD blocks, author byline, dateModified, internal links. Anything missing is invisible to Perplexity.

5.2 Primary Source Attribution (HIGHEST leverage, Perplexity specific)

The single signal that most distinguishes Perplexity from ChatGPT Search. Per the Mike Khorev 2026 Perplexity SEO guide, the Wellows 2025 visibility study, and the Authority Tech 2026 strategy analysis, Perplexity rewards source pages that themselves cite primary sources. The retrieval ranker treats well cited content as more trustworthy.

Standard:

• Every factual claim that is not common knowledge has an inline outbound link to a primary source: government data, academic study, peer reviewed paper, regulatory filing, original press release, named expert
• Approximately one outbound link to a .gov, .edu, or peer reviewed journal per 200 to 300 words of substantive content
• Citations are inline HTML links, not footnotes (Perplexity's retrieval reader weights inline link density per the Wellows 2025 study)
• Anchor text names the source and year ("per the NIH 2024 obesity prevalence report") so the model can extract the attribution

Per Internet Search Inc's 2025 ranking factors analysis, pages with 5 or more authoritative outbound links per 1,000 words earned approximately 2.4 times the citation rate of pages with fewer than 2.

5.3 Freshness and Visible Dates (HIGHEST leverage, Perplexity specific)

Per the Wellows 2025 visibility study, content older than 30 days dropped approximately 40 percent in citation rate; older than 90 days dropped approximately 65 percent. Per the Mike Khorev 2026 ranking analysis, content updated within the past 2 to 3 days gets a visibility boost on time sensitive topics.

Standard:

• dateModified visible in UI and in schema
• Both datePublished and dateModified in article header, not footer
• Genuine substantive update, not date manipulation (see framework-hcs.md)
• Time sensitive content on documented cadence (Section 14); evergreen content quarterly substantive review

Per the Otterly 2026 SEO guide, sites with quarterly cadences hold citation positions more reliably than sites with annual or ad hoc updates.

5.4 Citation Transparency Mirroring (HIGH leverage, Perplexity specific)

Perplexity surfaces sources prominently and treats sites that cite transparently as aligned with its retrieval philosophy. Per the Mike Khorev 2026 SEO analysis and the Skale 2026 GEO guide, pages with visible References sections and credentialed author bylines earn citation at approximately 1.7 to 2.1 times the rate of pages with implicit attribution.

Standard: visible References section listing primary sources with author, year, publication; author byline linking to a credentialed bio (see framework-eeat.md); reviewer credit on YMYL content (see framework-ymyl.md); cite attributes on <blockquote> for expert quotes.

5.5 Direct Answer Density and Section Length (HIGH leverage, Perplexity specific)

Per the Wellows 2025 study and the Discovered Labs 2025 to 2026 analysis, Perplexity prefers high factual claim density. Standard: lede 40 to 75 words, direct answer; every H2 opens with 40 to 75 word answer; sections of 120 to 180 words earn higher citation rates than 50 word fragments or 400 word walls; one citable hyperlinked statistic per 150 to 200 words.

5.6 Schema Completeness (HIGH leverage, Shared)

Server rendered JSON-LD with @graph pattern, @id cross references, and sameAs links earns higher citation rates on every AI surface per the Discovered Labs 2025 to 2026 analyses. Stack: Organization (sameAs Wikipedia, Wikidata, LinkedIn, Crunchbase), WebSite, WebPage, Article or BlogPosting (author, datePublished, dateModified), FAQPage mirroring visible <details>, BreadcrumbList, Person (sameAs ORCID, LinkedIn). See framework-schema.md. Critical for Perplexity: Article dateModified and visible page date must match; mismatches are discounted.

5.7 Entity Recognition (HIGH leverage, Perplexity specific)

Per the Wellows 2025 study and the Skale 2026 GEO guide, Perplexity uses entity recognition to match query entities to candidate pages. A reader new to the site should identify business name, owner, location, and topic in the first 200 words. Schema sameAs connects to Wikipedia, Wikidata Q-ID (see framework-knowledgegraph.md), LinkedIn, industry registries. Per the Authority Tech 2026 analysis, brands with stable entity declarations earn approximately 1.9 times more citations on branded queries.

5.8 Academic and Government Source Preference (HIGH leverage, Perplexity specific)

Per the Skywork AI 2025 accuracy tests, the Jenni AI 2025 academic research analysis, and the Truescho 2026 Perplexity Pro research guide, Perplexity weights .edu, .gov, .org (research and standards bodies), and peer reviewed journal domains higher than other sources. Perplexity Pro's Academic Focus mode limits search to peer reviewed sources via Semantic Scholar (approximately 200 million papers); arXiv, bioRxiv, SSRN feature prominently. In May 2025 Perplexity partnered with Wiley per the Skywork 2025 sources breakdown.

Standard for research positioning: outbound links to peer reviewed primary sources; author credentials displayed; [Author Year] style inline citations; deposit preprints in the relevant repository (arXiv for STEM, SSRN for social science, bioRxiv for life sciences).

Per the Skywork 2025 tests, Perplexity exhibits a 5 to 10 percent hallucination rate on academic citations on Pro, so clients with open access primary sources earn disproportionate citation share.

5.9 Reddit, Forum, and Community Signals (MEDIUM leverage, Perplexity specific)

Per the Profound August 2024 to June 2025 analysis, the Otterly January to February 2026 analysis (1 million plus citations), and the Tech Edge AI 2026 citation index, Reddit is the largest external source category in Perplexity citations. On community queries Reddit reached approximately 46.7 percent of citations per Discovered Labs 2025 to 2026; approximately 24 percent across all categories by January 2026 per Tech Edge AI 2026.

For clients with community presence: genuine subreddit participation; long form comments with primary source links; brand mentions in discussion threads; avoid karma manipulation, paid promotion, AI generated comments.

Caution: per the PR Newswire August 2025 5W index, a Google parameter change caused ChatGPT's Reddit share to swing from 60 to 10 percent in six weeks; Perplexity showed smaller but real volatility. Over rotation onto one platform is dangerous. See framework-knowledgegraph.md and framework-trustsignals.md.

5.10 YouTube Transcript Mentions (MEDIUM leverage, Perplexity specific)

Per the Ahrefs December 2025 correlation study cited in the position.digital 2026 AI SEO statistics roundup, YouTube mentions correlate with AI search visibility at 0.737, the strongest single predictor measured. Perplexity ingests YouTube transcripts; brand mentions in popular videos translate to citation probability.

5.11 Google Index Adjacency (MEDIUM leverage, Shared)

Per the Skale 2026 GEO guide, Surfer SEO 2026 ranking guide, and PragoMedia March 2026 analysis, Perplexity often references top ranking Google pages even though it does not use Google's index directly. Position 1 to 3 Google rankings correlate strongly with Perplexity citation for the same query. See framework-technicalseo.md, framework-internallinking.md, framework-keywordresearch.md.

5.12 Brand Mention Pace (MEDIUM leverage, Perplexity specific)

Per the BrightEdge 2025 to 2026 guide, the SEMrush 2026 most cited domains analysis, and the Otterly 2026 SEO guide, brand mentions across industry publications, podcasts, review platforms, forums, and YouTube compound Perplexity visibility. Standard: 5 to 10 new third party mentions per quarter on industry, regulatory, or academic sources; tracked in a mention monitoring tool (see framework-digitalpr.md). Mentions on .gov, .edu, recognized industry publications, and Wikipedia weight highest.

5.13 Signal Triage Summary

Highest leverage for new engagements: 5.1 Substrate, 5.2 Primary Source Attribution, 5.3 Freshness, 5.6 Schema, 5.7 Entity Recognition, 5.8 Academic and Government Preference. These six gate citation eligibility. Medium leverage (5.4, 5.5, 5.9, 5.10, 5.11, 5.12) compound and stabilize against volatility.

6. Citation Patterns on Perplexity

6.1 Sources Panel Mechanics

Every Perplexity Search response surfaces a Sources panel inline with the answer. Per the BrightEdge 2025 to 2026 guide, the average response cites approximately 8.79 sources; per Discovered Labs 2025 to 2026, the per response count ranges from 5 to 15. This is dramatically higher than ChatGPT Search's 3 to 7 (see framework-searchgpt.md Section 7.1) and Google AI Overview's 3 to 8 (see framework-aioverviews.md). The product was built around prominent citation as a trust signal.

6.2 Citation Selection Criteria

Inferred from 2025 to 2026 studies: source in the retrieval candidate set (PerplexityBot index plus Perplexity-User access); source contributes a unique factual element; source readable in reading mode; source cites primary sources itself (Section 5.2); source has recent dateModified (Section 5.3); source meets E-E-A-T confidence threshold; source connects to a recognized entity graph. Pages failing the first three are not retrieved.

6.3 Citation Position Within The Sources Panel

The panel orders by retrieval relevance, not chronologically. Per the Otterly 2026 SEO guide and BrightEdge 2025 to 2026 guide, the top three sources earn approximately 60 to 70 percent of click through.

6.4 Why Perplexity Cites More Transparently Than Other Engines

Per the Demand Local 2026 ROI statistics and the Otterly 2026 SEO analysis, Perplexity surfaces citations on every answer by default. ChatGPT Search hides them behind a Sources affordance; Google AI Overview shows chips requiring click. Consequences: cited domains in Perplexity earn approximately 2 to 4 times the click through of cited domains in ChatGPT Search per Demand Local 2026; Perplexity is the most measurable AI citation channel because every cited domain is observable per Otterly 2026; public visibility creates accuracy accountability the other engines do not face equally (Perplexity hallucination rate approximately 5 to 10 percent on Pro per the Skywork 2025 accuracy tests).

6.5 Citation Persistence Across Reruns

Per the BrightEdge volatility study, 96.8 percent of cited domains and 97.2 percent of mentioned brands showed no weekly change. Of the 3.2 percent that did change, approximately 87 percent were declines. Run tracking weekly; investigate consistent declines across multiple samples.

6.6 Brand Mention vs Citation

A mention names the brand without a cited source; a citation includes the domain in the Sources panel. Perplexity's mention to citation ratio is approximately 1.4 to 1 per the BrightEdge 2025 to 2026 analysis, against ChatGPT's approximately 3.2 to 1. Citation carries the primary KPI weight on Perplexity.

7. Optimizing for Perplexity Pages

7.1 What Perplexity Pages Are

Per the Pages launch blog (May 30, 2024), VentureBeat May 2024 coverage, the Writingmate explainer, and the Search Engine Journal December 2024 Discover Pages analysis, Pages is a long form publishing surface. A user submits a topic prompt; Perplexity drafts a structured article with title, subheadings, media, and inline citations; the user edits and publishes. Published Pages live at https://www.perplexity.ai/page/[slug]-[id] and become Google indexable.

This is a third party publication surface. The site cannot push existing content into Pages. The site can be cited within a Page, have a client team author a Page citing the client site, or treat Pages as a digital PR target.

7.2 When Pages Surface in Search Results

Per the Search Engine Journal December 2024 Discover Pages analysis, Pages appear in Perplexity Search results for related queries and in Google when indexed (usually hours to days). For long tail topics, a competing Page can outrank the client for the client's own topic.

Two responses: Defensive Pages (author a Page citing the client site to claim the surface first); Offensive Pages (extend visibility into Perplexity's user generated content library).

7.3 Authoring a Page from a Client Account

Workflow (Pro account): open Perplexity, navigate to Pages; submit a topic prompt aligned with client authority; select target audience; review the draft and verify cited sources include the client site; edit headings, body, and inline citations for accuracy; add or replace media; publish; capture the URL. Submit to Google Search Console via URL Inspection per framework-technicalseo.md.

7.4 The Page Content Pattern

Per the Skim AI 2024 Pages analysis and the Generative AI Pub 2024 analysis, a high performing Page follows:

7.5 Pages and Site Citation Strategy

A Page that cites a client article surfaces both. The cited article must pass the Section 5 audit; the Page should be authored from an account associated with the client; the Page should be promoted (LinkedIn, Reddit, internal site link) for backlinks; revisit quarterly.

7.6 What Not To Do

Per the Skim AI 2024 analysis and the Liaxinweng 2024 Pages failure analysis: no editorial review (unedited output hallucinates sources); backlink mill Pages with no informational value decline; Pages are not a substitute for first party site content; head terms where the client cannot compete fail (Pages perform better on long tail).

8. Perplexity Spaces

8.1 What Spaces Are

Per the Perplexity Internal Knowledge Search and Spaces announcement (October to November 2024), the Gilbane October 2024 coverage, and the Beginners in AI 2026 guide, Spaces are persistent customizable research workspaces with three distinguishing features:

1. Custom system prompt. Free text instructions defining how Perplexity searches, what sources to prioritize, citation format, output structure. Applies to every thread.
2. Uploaded files. Pro supports up to 50 (PDFs, Word, Excel, PowerPoint, CSV); Enterprise up to 500. The AI searches uploaded files and the open web in one thread.
3. Connectors and collaboration. Pro and Enterprise connect SharePoint, OneDrive, Google Drive; Enterprise invites viewer or contributor; threads accumulate shared context.

Over 5 million Spaces existed by Q1 2026 per Panto AI 2026; Spaces are Perplexity's fastest growing surface.

8.2 How Site Content Surfaces in a Space

Space contents are private context for that Space's assistant. Site content surfaces in someone else's Space only when that user runs a web search retrieving the client site. The site cannot push content into Spaces. Section 5 signals matter inside Spaces the same way as in standard Perplexity Search; no separate Spaces optimization.

8.3 When To Build A Space For An Organizational Client

A Space is the client's internal tool, not a client side citation surface. Build when the team needs a persistent AI workspace tuned to their domain; the client centralizes internal documents searchable alongside the open web; the client standardizes methodology (system prompt enforces it); the client onboards team members into a research workflow. Operational work, not SEO, but fits within the same engagement for research heavy clients.

8.4 Space System Prompt Patterns

Per the Understanding AI 2026 master prompt guide, the GitHub TheBigPromptLibrary collection, and the Perplexity Prompt Guide documentation, an effective system prompt specifies tone and audience, source priority (peer reviewed, government, primary documents), citation format ("[Author Year] inline; Sources section"), output structure (60 to 100 word executive summary, 150 to 200 word sections), and negative constraints ("Do not infer beyond retrieved sources. Flag uncertainty."). For regulated industries the system prompt documents the research process for compliance review.

8.5 Discoverability of Public Spaces

Public Spaces surface in Perplexity's Discover and Spaces explore views and drive traffic to underlying sources. For teams willing to publish, this is a digital PR surface adjacent to Pages (Section 7).

9. Perplexity Comet Browser

9.1 What Comet Is

Per the Perplexity Comet launch blog (July 9, 2025), the IBM Think July 2025 analysis, and the No Hacks 2026 agentic browser landscape, Comet is a Chromium browser with Perplexity's agent integrated into the chrome. The user interacts via address bar, side panel, and voice; the agent reads tabs, fills forms, compares products, executes transactions.

Trajectory per Wikipedia: desktop Max tier (200 USD per month) July 2025; Android November 2025; iOS free March 2026. Per SEO Sandwich 2026, approximately 3 million MAU by early 2026, DAU up 320 percent YoY, approximately 1.9 percent global browser share.

9.2 Why Comet Shifts SEO

Per the Pravin Kumar 2026 Webflow impact analysis, the IBM Think 2025 analysis, and framework-agenticaisearch.md, agent browsers shift the citation surface. When Comet acts on behalf of the user, the user does not always see source pages. The source earns an agentic citation: read and acted on even when the user never clicks through.

Consequences: substrate matters more (the agent's reasoning often reads the DOM directly even though Comet renders JS); machine readable commerce data wins (Product/Offer schema, pricing tables, comparison matrices); CAPTCHA and hard friction lose (agent browsers deprioritize gated pages per IBM Think 2025; OpenAI's Operator was deprecated partly for JS heavy CAPTCHA failures per TechCrunch August 2025); brand consistency matters in agent memory (Comet maintains browsing memory; consistent NAP and entity declarations earn higher recall).

9.3 Optimization For Comet Specifically

Beyond Section 5: Product, Service, Offer, FAQPage, HowTo schema (see framework-schema.md); unambiguous CTAs with reachable destination URLs; no CAPTCHA on critical commercial pages; crisp prices, hours, availability in visible HTML and matching schema; stable URLs (agent memory references URLs; broken redirects degrade confidence).

9.4 How Comet Traffic Identifies Itself

Comet sessions identify as Chromium to the destination per the Comet feature documentation. Automated actions sometimes route through Perplexity-User per the 2026 Comet reviews. Server logs should treat Comet as ordinary Chromium. Optimize substrate and structured data so both agent and human visitors find what they need.

9.5 Agentic Browser Landscape Context

Comet shares architecture with ChatGPT Atlas (October 21, 2025, see framework-searchgpt.md Section 11). Brave added agentic features in 2025; Dia (The Browser Company) is in private beta per the No Hacks 2026 landscape. Optimizing for Comet optimizes for Atlas and the broader category.

9.6 The Long Tail Of Browser Memory

Comet builds long term memory of visited sites per the Comet documentation and the Make Use Of 2026 long term review. Sites with consistent brand identity and stable URLs earn higher recall. Slow signal, defensible moat for established businesses.

10. llms.txt and AI Surface Maps for Perplexity

10.1 What llms.txt Is

llms.txt is a proposed standard from Jeremy Howard of Answer.AI in September 2024. A Markdown file at /llms.txt that provides AI engines a curated guide to the site's most important content, with optional /llms-full.txt expanded version. See Section 9 of framework-searchgpt.md for the foundational treatment.

10.2 Current Adoption State (Q2 2026)

Per the SERanking November 2025 research on approximately 300,000 domains and the BuiltWith October 25, 2025 tracking, approximately 10.13 percent of sites have llms.txt; approximately 844,000 websites globally as of late 2025.

Per the Studio Ubique 2025 analysis, the Webless 2026 AI bots crawl rules analysis, the Goodie 2025 LLMs.txt and Robots.txt analysis, and the Big Cloudy 2026 guide, Perplexity has publicly stated support for the llms.txt standard. This positions Perplexity ahead of OpenAI and Google on llms.txt adoption commitment.

10.3 The Framework Position For Perplexity Sites

Given Perplexity's stated support, the framework recommends llms.txt as medium priority for Perplexity optimization (against low priority hygiene for OpenAI work per framework-searchgpt.md Section 9.3). Cost low, maintenance minimal (quarterly review).

10.4 Recommended Structure For Perplexity Optimization

# Example Business

> One sentence description of the business, its primary topic authority, and what content the site offers.

## Core Pages

- [About](/about/): who we are, credentials, location
- [Authors](/authors/): expert team, Person schema, sameAs network
- [Methodology](/methodology/): research and editorial process

## Authoritative Content

- [Cornerstone topic 1](/topic-1/): definitive guide with primary source citations
- [Cornerstone topic 2](/topic-2/): peer reviewed references

## Recent Updates

- [Latest article](/blog/latest/): published 2026-MM-DD
- [Methodology update](/methodology/): updated 2026-MM-DD

## Primary Sources We Cite

- Government data: primary `.gov` references
- Academic sources: journals, arXiv categories, institutional repositories
- Industry standards: standards bodies, regulators

## Schema and Entity

- Wikipedia, Wikidata Q-ID, LinkedIn, ORCID

The Perplexity specific elements are the Primary Sources We Cite section (Section 5.2) and the explicit Methodology link (Section 5.4).

10.5 Validating The File

/llms.txt should return HTTP 200 with Content-Type text/plain or text/markdown, be at the root path, validate as Markdown, reflect current site structure (reviewed quarterly), and not be blocked in robots.txt.

curl -A "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; PerplexityBot/1.0; +https://perplexity.ai/perplexitybot)" \
  -s -o /dev/null -w "%{http_code} %{content_type}\n" https://example.com/llms.txt

11. Cross Engine Differentiation

Why optimizing for Perplexity is different from optimizing for ChatGPT Search or Google AI Overviews. This section reads as a summary; treat it as the executive briefing for client conversations where the question "we already do AI SEO, isn't this the same?" arises.

11.1 Source Preference Matrix

The Perplexity bimodal pattern (Reddit plus primary sources) is the central differentiator. Optimization must support both ends.

11.2 Citation Density Per Response

11.3 Freshness Sensitivity

Perplexity's steep decay means a quarterly refresh cadence is the minimum; monthly is preferred for time sensitive verticals.

11.4 Reading Mode Strictness

All three surfaces require server rendered substrate. Perplexity and ChatGPT bots equivalently strict (no JavaScript per Section 5.1). Googlebot renders JavaScript with delay. Optimize for the strictest floor (Perplexity and ChatGPT) and Google AI Overview optimization comes free.

11.5 Brand Mention To Citation Ratio

Perplexity's tight ratio means citations carry primary KPI weight; track mention rate as secondary.

11.6 Optimization Sequencing

For sites targeting all three surfaces:

1. Substrate. Apply framework-contentfirst.md.
2. Perplexity. This framework Sections 4 through 6, 10.
3. ChatGPT Search. framework-searchgpt.md.
4. Google AI Overview. framework-aioverviews.md.

Perplexity is sequenced first because its primary source attribution signals are the most distinctive; sites earning them tend to earn ChatGPT and AI Overview citation downstream. The reverse is not as reliably true.

12. Common Perplexity SEO Mistakes (Top 10 Anti Patterns)

The ten anti patterns below cause silent Perplexity invisibility. Addressing any three on a site that previously failed produces measurable citation lift within four to eight weeks.

12.1 Blocking Perplexity-User While Allowing PerplexityBot

Most common invisibility cause for sites that intended visibility. Site appears in Perplexity index but is not re fetched during user queries; freshness stales and citation rate degrades. Fix per Section 4.4.

12.2 No Primary Source Citations on the Source Page

The signal that most distinguishes Perplexity ranking. Sites that make claims without inline primary source links are systematically discounted per Section 5.2.

12.3 Date Drift Between Schema and Visible Page

Schema says dateModified is 2026-04-15; visible footer says "Last updated 2024". Perplexity detects and discounts. Fix: schema and visible date must match and reflect genuine updates.

12.4 Client Side Rendered Content

React or Vue SPA delivers an empty root div. Perplexity bots see nothing. Fix per framework-contentfirst.md and Section 5.1.

12.5 No Visible References Section

Per Section 5.4, sites with no References section earn approximately half the citation rate of sites with explicit attribution. Fix: References section listing primary sources with author, year, publication.

12.6 Schema Injected Via GTM Or Client Side JS

Perplexity bots see no schema. Fix: server rendered JSON-LD in document head, validated against Schema.org and Google's Rich Results Test.

12.7 Old Content Never Refreshed

The 40 percent citation drop beyond 30 days and 65 percent beyond 90 days per the Wellows 2025 study makes freshness the second highest leverage signal. Fix per Section 14 cadence.

12.8 Brand Entity Ambiguity

Multiple businesses share the name, or inconsistent naming across the web. Perplexity cannot resolve which entity. Fix per framework-entitysalience.md and framework-knowledgegraph.md.

12.9 Reddit Posture Absent For Community Topics

For consumer products, lifestyle, recreation, hobbyist tools, sites with no Reddit posture lose to sites whose brand appears in relevant subreddits per Section 5.9. Fix: organic subreddit engagement, no karma manipulation.

12.10 Treating Perplexity As "Same As ChatGPT"

The single most expensive strategic mistake. Per Section 11, the engines cite different sources, weight different signals, and reward different content. Applying ChatGPT tactics (Wikipedia heavy positioning, Bing IndexNow focus) and expecting Perplexity citation typically gets neither.

13. Audit Rubric

13.1 First 90 Days Page Subset Rubric

The five item gating rubric. Every priority page must pass these five for citation consideration.

Pages passing these five are in the candidate pool; failing any disqualifies.

13.2 Per Page Audit Rubric (Full)

Score 15. World class: 13+ with zero F1 to F5 fails.

13.3 Site Wide Audit Rubric

Score 15. World class: 13+ with zero critical fails on S1 through S4.

13.4 First 90 Days Tracking

13.5 Perplexity Specific Citation Tracking

Manual sampling protocol:

1. Open Perplexity Search
2. Submit priority query
3. Inspect Sources panel inline with answer
4. Record cited domains and Sources panel order
5. Record mention vs citation status (Section 6.6)
6. Record citation count (benchmark approximately 8.79 per BrightEdge)
7. Repeat weekly across the priority query set

Tooling: Otterly is the lowest cost dedicated Perplexity tracker; Profound and AthenaHQ provide enterprise scale tracking. Manual sampling is feasible for sub 50 query lists.

13.6 Page Tracking (Clients With Pages Authored)

13.7 Space Tracking (Clients With Spaces Built)

14. Maintenance Schedule and Report Templates

14.1 Maintenance Cadence

Weekly. Sample top 10 priority queries; record citation state and Sources panel position. Check server logs for PerplexityBot and Perplexity-User on priority pages. Refresh one priority page substantively.

Monthly. Sample next 50 priority queries. Schema validation sweep on priority pages. Review new competitor candidates. Refresh time sensitive content. Audit robots.txt. Review client owned Pages.

Quarterly. Site wide Section 13 audit. Refresh evergreen content. Review brand mention pace. Update llms.txt. Refresh Space system prompts and files. Audit Reddit posture.

Annually. Full framework review against current Perplexity surface state (Comet was new July 2025; Spaces October 2024; Pages May 2024). Refresh underlying frameworks. Strategic review of priority query selection.

14.2 Implementation Report Template

# Perplexity Framework Implementation Report

**Site**: {{BUSINESS_NAME}}
**Date**: {{TODAY}}

## Summary
- Priority pages audited / passing F1-F5 / passing full rubric
- Site wide score: X/15

## Baseline Citation State
- Queries sampled, cited, average count, mention rate, bot fetches last 30 days

## Substrate
- Pages migrated server rendered; schema moved out of JS; curl pass before/after

## Bot Access
- Robots.txt corrections; log verification of PerplexityBot and Perplexity-User

## Primary Source Citations
- Outbound links added by category; References sections added

## Schema, Entity, Freshness
- Counts per category

## llms.txt and Pages/Spaces (where applicable)
- File created; Pages authored with URLs; Spaces built with system prompt

## Tracking
- Tracker selected, cadence documented, log monitoring deployed

## Sign-Off

14.3 Audit Report Template

# Perplexity Framework Audit Report

**Site**: {{BUSINESS_NAME}}
**Date**: {{TODAY}}

## Executive Summary
One paragraph assessment.

**Site score**: X/15
**Page average**: X/15
**Citation rate** (weekly sampled): X percent
**Avg citation count when cited** (benchmark 8.79): X
**Mention rate**: X percent

## Findings By Section
Substrate, bot access, primary sources, schema, entity, freshness, Reddit, mentions

## Critical Failures
List with remediation

## First 90 Days Subset Findings
F1 to F5 table per priority page

## Recommended Remediation Order
Critical: substrate, bot access, primary source citations
High: schema, entity, freshness, References sections
Medium: content density, internal linking, llms.txt, Reddit, Pages

## Sign-Off

14.4 Server Log Monitoring Bash Script

#!/usr/bin/env bash
# /var/www/sites/[domain]/scripts/perplexity-bot-monthly-report.sh

LOG="/var/log/nginx/access.log"
ROLL="/var/log/nginx/access.log.1"
DOMAIN="$1"
DATE=$(date +%Y-%m-%d)
DIR="/var/www/sites/${DOMAIN}/reports/perplexity-bots"
mkdir -p "${DIR}"
OUT="${DIR}/${DATE}.txt"

{
  echo "Perplexity Bot Access Report - ${DOMAIN} - ${DATE}"
  echo ""
  echo "PerplexityBot fetches:"
  grep -c "PerplexityBot" "${LOG}" "${ROLL}" 2>/dev/null
  echo "Perplexity-User fetches:"
  grep -c "Perplexity-User" "${LOG}" "${ROLL}" 2>/dev/null
  echo ""
  echo "Top paths by PerplexityBot:"
  grep "PerplexityBot" "${LOG}" "${ROLL}" 2>/dev/null | \
    awk '{print $7}' | sort | uniq -c | sort -rn | head -25
  echo ""
  echo "Non 200 status codes (potential blocks):"
  grep -E "PerplexityBot|Perplexity-User" "${LOG}" "${ROLL}" 2>/dev/null | \
    awk '$9 != 200 {print $9, $7}' | sort | uniq -c | sort -rn | head -25
  echo ""
  echo "Generic Chrome on macOS UA from non residential IPs (undeclared crawler signal):"
  grep "Macintosh; Intel Mac OS X" "${LOG}" "${ROLL}" 2>/dev/null | \
    awk '{print $1}' | sort | uniq -c | sort -rn | head -25
} > "${OUT}"

echo "Report at ${OUT}"

Save at /var/www/sites/[domain]/scripts/perplexity-bot-monthly-report.sh, chmod +x, cron monthly.

14.5 Quarterly Citation Comparison Worksheet

For each priority query, sampled weekly and aggregated quarterly:

Use to spot engines diverging (engine specific algorithm change), universal declines (content quality or freshness drift), universal gains (substrate or entity work paying off across surfaces).

End of Framework Document

Document version: 1.0 Created: 2026-05-14 Maintained by: ThatDeveloperGuy

Perplexity Search, Pages, Spaces, and Comet constitute a distinct AI surface family with approximately 45 million MAU and 1.2 to 1.5 billion queries per month by mid 2026. The signals driving Perplexity citation are reading mode survival, primary source attribution on the source page, citation transparency, freshness within tight windows, schema completeness, entity recognition, and academic source preference. The most common invisibility cause is blocking Perplexity-User while allowing PerplexityBot; the second most common is failing to cite primary sources on the source page itself.

Perplexity is the most measurable AI citation channel and the most tightly bound to primary source attribution. Apply after the substrate doctrine in framework-contentfirst.md, alongside framework-searchgpt.md, framework-aioverviews.md, and framework-aicitations.md.

Companions

• framework-contentfirst.md, substrate doctrine, prerequisite
• framework-aicitations.md, broader AI citation
• framework-aioverviews.md, Google AI Overviews and AI Mode
• framework-searchgpt.md, OpenAI surface family
• framework-hcs.md, framework-infogain.md, framework-eeat.md, framework-sqrg.md, framework-ymyl.md
• framework-schema.md, framework-entitysalience.md, framework-knowledgegraph.md
• framework-agenticaisearch.md, agentic surface (Comet, Atlas)
• framework-multimodalsearch.md, framework-internallinking.md
• framework-cross-stack-implementation.md, framework-react.md, framework-tailwind.md
• framework-digitalpr.md, framework-trustsignals.md
• SEO-Search-Appearance.md, surface map, four pillars
• SERP-Optimization.md, feature targeting
• Scheduled Phase 2: framework-hreflang, framework-robotstxt-crawlbudget, framework-brandvoice, framework-ai-content-workflow, framework-ppc-seo-coordination, framework-aso, framework-contentbriefs
• 14 tier Engine Optimization Stack, Tier 3 AI Search Domination