> ## Documentation Index
> Fetch the complete documentation index at: https://docs.anymorph.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Perplexity

> Citation-first AI search reference engine

## Background: provider and product surface

Perplexity is an AI-native search engine from Perplexity AI. Its user base is smaller than ChatGPT or Google, but its citation transparency makes it especially useful for GEO experimentation.

## Usage scale and market context

| Metric                 | Reported value                              | Context                                                     |
| ---------------------- | ------------------------------------------- | ----------------------------------------------------------- |
| Query volume           | \~100M queries/month                        | Frequently cited industry estimate; varies by source/date   |
| AI search market share | Single-digit share                          | Estimated below ChatGPT/Google but above many niche engines |
| Product context        | Smaller user base, high source transparency | Lower MAU, stronger citation observability                  |

## Product surface and user experience

* Citation-first UI with inline citations attached to answer claims.
* Pro/Deep Research reads more sources and performs iterative retrieval.
* Search, answer synthesis, and citation are tightly coupled.
* Useful for source comparison and citation auditing in SaaS/product/research prompts.

## GEO characteristics

| Axis                | Engine-specific behavior                                          | GEO interpretation                                                                      |
| ------------------- | ----------------------------------------------------------------- | --------------------------------------------------------------------------------------- |
| Retrieval substrate | Own web-scale index + hybrid retrieval + reranking                | Does not simply copy Google/Bing ranking                                                |
| Citation UI         | Citation-first inline source UI                                   | Best observability for source strategy experiments                                      |
| Query fan-out       | Multi-step search plans in Pro/Deep Research                      | Sequentially explores vendor/review/pricing/source branches                             |
| Source diversity    | High unique-domain count and average citations are often observed | Diverse evidence layer matters                                                          |
| Measurement note    | Separate source-heavy evidence layer from final answer surface    | homepage inline share 55.1%, pricing/plans 50.3%; blogs/news/reports are evidence layer |

### Key cited / influential source types

| Source type                        | Observed pattern                                                                                      | GEO action                                              |
| ---------------------------------- | ----------------------------------------------------------------------------------------------------- | ------------------------------------------------------- |
| Reddit / community                 | Reddit is highly prominent in Perplexity source patterns                                              | Manage real-user, complaint, and alternative narratives |
| YouTube                            | YouTube appears in Perplexity top-source groups                                                       | Use as how-to/demo/explainer source                     |
| News / earned media                | Perplexity often skews toward earned/news sources                                                     | Secure PR and industry media coverage                   |
| Industry authority / analyst sites | Analyst and niche expert sources                                                                      | Important for B2B comparisons                           |
| Owned reference pages              | homepage inline share 55.1%, pricing/plans 50.3%; guide/pricing/product pages are citation candidates | Use BLUF, dates, tables, and schema                     |

## Engine-specific GEO strategy

| Strategy                            | Action                                                          | Metric                  |
| ----------------------------------- | --------------------------------------------------------------- | ----------------------- |
| Use as source strategy test engine  | Quickly observe which source types survive                      | Inline citation count   |
| Create BLUF/self-contained passages | Put answer paragraphs and comparison tables under H2s           | Passage-level citation  |
| Make freshness explicit             | Show last updated, publication date, and recent numbers         | Fresh source inclusion  |
| Design third-party source mix       | Manage Reddit/YouTube/news/review/authority sources per cluster | Source domain diversity |

## What to measure

* Usage scale and surface-specific exposure
* Source/citation presence and source type
* Brand mention, recommendation, sentiment, and claim accuracy
* Engine-specific failure mode: retrieval, citation, or recommendation

## Sources

* [https://ziptie.dev/blog/how-perplexity-ai-answers-work/](https://ziptie.dev/blog/how-perplexity-ai-answers-work/)
* [https://research.perplexity.ai/articles/architecting-and-evaluating-an-ai-first-search-api](https://research.perplexity.ai/articles/architecting-and-evaluating-an-ai-first-search-api)
* [https://www.stackmatix.com/blog/perplexity-ai-optimization-strategy](https://www.stackmatix.com/blog/perplexity-ai-optimization-strategy)