A quick field guide
From a check
to a better page.
Start an AI search review with evidence you can inspect. This free AI search readiness checker looks at a page’s available HTML, declared access rules and observable source signals. It helps organize the next technical or editorial task, while keeping actual mentions and citations as a separate research question.
- 01
Choose a page and source
Fetch a public URL, or paste HTML and optional robots.txt. Add a business name for a literal text check.
- 02
Follow three evidence groups
Review access, readable information and source signals. Open findings to understand their limits.
- 03
Choose the next task
Correct an unintended restriction, clarify the information or plan a separate observation of real AI answers.
A closer look / 05
Observation first. Judgment second.
HTML, directives and declared rules
Check purpose, facts and missing context
A correction, a source review or dated answer sampling
Readiness and visibility are different questions
A readiness review asks what a page makes available and how clearly it describes its subject. A visibility study asks whether a particular system mentioned or cited it in an observed answer. This tool handles the first job. It does not query ChatGPT, Gemini, Claude, Perplexity or Google’s AI features, and it does not turn a checklist into a citation probability.
Observed counts describe checks in this report. They are not an AI-provider score or a competition ranking. A page with no review flags may still need better information, stronger original evidence or a more relevant answer. A page with a missing optional element may already be useful. Read the finding, its source and its next step together.
Follow the evidence through three checks
Access signals include the fetched HTTP response, available indexing and snippet directives, and declared robots.txt rules when supplied. A robots simulation shows matching rules for the checked path. It is not a request from the named crawler and cannot verify a firewall, provider crawl, search index or later policy change. Missing or failed robots retrieval remains visibly unassessed.
Readable information means text, headings and links found in the supplied HTML. JavaScript is not executed. If the meaningful answer is absent from the fetched document, inspect the rendered page and actual crawler evidence before deciding what any specific system can use. Lists or question headings are not given automatic points.
Source signals are narrow observations such as an author field, an About-like link, a contact path or the optional literal business name. These do not verify expertise, truth or reputation. Open the referenced pages, inspect the evidence behind claims and decide whether attribution is appropriate for the content. The GEO & AI Search learning path helps turn those observations into a transparent research plan.
Keep search access and training choices separate
Google documents Googlebot as the crawl control for Google Search, including AI Overviews and AI Mode. Google-Extended serves different purposes in other Google AI systems. OpenAI separately documents OAI-SearchBot for search and GPTBot for potential training use. PerplexityBot is documented for surfacing and linking websites in Perplexity search.
The table separates these purposes and shows the matching declared rule. Allowing a training crawler is a business decision, not a way to improve a score here. User-triggered fetchers have distinct behavior and are excluded from this robots simulation. Check the provider documentation and your own access policy before changing production rules.
Google also says its AI search features have no additional technical requirements beyond the relevant search foundations. Special AI files, llms.txt or a special schema type are not required for those features. That statement concerns Google’s documented features; it should not be generalized into a claim about every retrieval system.
Choose the next useful GEO task
Fix unintended restrictions first if the page should be publicly available. Next examine whether the main information is delivered clearly and supported by real sources. Record what changed, the page tested and the date. A meaningful follow-up might be checking a factual claim or making authorship easier to understand, rather than adding more markup to satisfy a counter.
To study actual visibility, define a small set of real audience questions, record the system and wording, save dated answers and inspect the sources. Keep that evidence separate from this technical report. For help connecting the checks with content and observation, explore our Generative Engine Optimization approach.
Worth knowing
Your questions,
answered.
Can this tell me whether ChatGPT cites my website?+
No. No AI answers are requested or sampled. This report describes available source and access signals, with explicit limits. Actual citation research needs observed answers and their context.
Do I need llms.txt for Google AI Overviews?+
Google’s current documentation says no new AI text file or special markup is needed for AI Overviews or AI Mode. This tool does not award points for llms.txt.
Should I allow every AI crawler?+
That depends on your access and content-use policy. Search discovery, training use and user-triggered requests serve different purposes. Review the current official documentation before changing rules.
Why are some checks marked not checked?+
The input may not include live headers, robots.txt or outside evidence. The report keeps unavailable data explicit instead of treating it as a pass or failure. Pasted HTML can still support useful source observations.