Free AEO checker · AI search, discovery and shopping agents

See your site the way AI agents see it

A free, reproducible grade of how Perplexity, ChatGPT and Gemini (Google grounding) read, cite and recommend your site. Fixes ranked by points gained.
New here? See the sample — a real audit, not a mock — then measure your own site with a free account.
48 checks16 crawlers in robots.txt7 fetched live3 answer enginesrubric v1.4.0free account
SAMPLE REPORT · A REAL AUDITprompeteer.ai
C+ 67 / 100
Retrievability
100
Parseability
40
Answerability
22
Entity and trust
83
Agent readiness
100
vs 5 competitors · measured 5 Sep 2026Open the sample →
The category

What is answer engine optimization?

Answer engine optimization is the practice of making a website readable, quotable and transactable by AI agents rather than only by search crawlers. It covers whether retrieval crawlers can fetch your pages, whether your content survives without JavaScript, whether a passage can be lifted out as a standalone answer, and whether a shopping agent can find your catalogue and complete a purchase.

Search

Can Perplexity, ChatGPT and Gemini (Google grounding) retrieve you — and do they cite you when a buyer asks?

Discovery

Can an agent find your machine-readable surfaces — markdown twins, well-known discovery, an MCP endpoint?

Commerce

Can a shopping agent read your catalogue and check out, against the ACP and UCP feed specifications?

How the grade is built

Five pillars, one weighted grade

48 deterministic checks. Every finding quotes the response it came from. A check we could not run is reported as unverified, never as a failure.
  • RetrievabilityDo robots.txt and your CDN admit 16 AI crawlers? 7 are fetched live.
    30%
  • ParseabilityHow much survives without JavaScript, which no AI crawler runs.
    25%
  • AnswerabilityCan a passage be lifted out and used as the answer?
    20%
  • Entity and trustDoes your schema resolve to a real organisation, author and date?
    15%
  • Agent readinessMarkdown twins, well-known discovery, MCP and product feeds.
    10%
Retrieval is not trainingBlocking GPTBot is a legitimate choice. Blocking OAI-SearchBot removes you from the answer. They are scored apart.
The crawler's view, not the browser'sWe fetch each page as a browser and as each crawler. A page that is empty to crawlers shows as the gap it is.
Schema must match the pageFAQ markup without matching visible questions is a penalty, not a credit.
Why now

Discovery is moving from search results to answers

Buyers increasingly ask an assistant instead of scanning results. When they do search, an AI answer now sits above the links. A page an AI crawler cannot read is absent from both.

WEEKLY CHATGPT USERS100M → 800M in 23 months
100MNov 23200MAug 24400MFeb 25800MOct 25
ChatGPT weekly active users
MonthUsers
Nov 2023100M
Aug 2024200M
Feb 2025400M
Oct 2025800M
OpenAI, reported figures, Nov 2023 – Oct 2025.
US GOOGLE SEARCHES ENDING WITHOUT A CLICK50.3% → 58.5%
50.3%201958.5%2024
US Google searches ending without a click
YearShare
201950.3%
202458.5%
SparkToro with Jumpshot (2019) and Datos (2024).
CLICKS TO THE #1 RESULT WHEN AN AI OVERVIEW APPEARS−34.5%
100no AI Overview65.5AI Overview shown
Indexed clicks to the top organic result
ConditionClicks (index)
No AI Overview100
AI Overview shown65.5
Ahrefs, 300,000 keywords, 2025.
TRADITIONAL SEARCH VOLUME BY 2026−25% (forecast)
100index, 2024752026 forecast
Traditional search engine volume, indexed (Gartner forecast)
YearVolume (index)
2024100
2026 (forecast)75
Gartner forecast, Feb 2024, for 2026 — outcome not yet published.
ADS INSIDE THE ANSWERSGoogle since 2024, ChatGPT since 2026
  1. Ads in AI Overviews (US)
  2. Google: AI Overviews monetising at about Search’s rate
  3. ChatGPT ads test (US): labelled Sponsored, below the answer
  4. ChatGPT ads in the UK, Mexico, Brazil, Japan and South Korea
Sources: OpenAI, Ads in ChatGPT · Alphabet Q3 2025 earnings call, 30 Oct 2025. Verified 5 Sep 2026.

Where 100 buying journeys are decided, at your assumption

Decided in an answer engineSearch, no clickSearch, click a result

25 of 100 journeys are decided where only a crawlable, citable page exists; 44 more end in search without a click. Only 31 reach a page at all.

Your assumption (the slider) × one published rate: 58.5% of searches end without a click (SparkToro/Datos, 2024). Not a forecast — a way to see the stakes.
Questions

What people ask before they run it

Is there a free AEO checker for my site?

Yes. This one runs 48 checks and queries 3 answer engines at no cost, on a free Prompeteer account. The report is yours to share, download and act on.

How do I make my blog visible to ChatGPT?

Serve the post body in the raw HTML and keep dateModified honest. Give every post a named author resolved in your schema graph, and open robots.txt to retrieval crawlers, not only training ones.

Do AI crawlers run JavaScript?

No major one does. Vercel and MERJ found no JavaScript execution across 500 million GPTBot fetches, and the same held for ClaudeBot and PerplexityBot. Googlebot remains the only major crawler that renders.

Does llms.txt help?

There is no evidence that it does. Google states it has no effect on Search, and no engine has confirmed reading it. We report whether you have one and weight it at zero rather than inflating your score with it.