CLUNKY AI
Method // ai-visibility-v1

AI visibility, scored without pretending to know a model’s mind.

The score measures observable public readiness signals. It does not claim to measure a universal “LLM ranking” and cannot predict whether a particular system will cite a site for a particular prompt.

Crawler access

25 points

Checks whether OAI-SearchBot, Googlebot and Bingbot are permitted to retrieve the scanned path and whether a sitemap is available. GPTBot training controls are deliberately not treated as ChatGPT Search visibility controls.

Entity clarity

25 points

Checks the title, summary, primary heading, relevant organisation or service structured data and the trail to About or Contact information.

Answer readiness

30 points

Checks crawlable detail, descriptive headings, direct question coverage, lists or tables and internal links that connect an answer to wider site context.

Public evidence

20 points

Checks for specific and verifiable proof markers, authorship or freshness information and supporting external sources. It does not decide whether a claim is true.

New scans

Direct signal checks

Scanner version 3.2 and later retrieves robots.txt and the sitemap, evaluates bot-specific rules for the scanned path and measures the visible page structure, entity markup and evidence signals. Unknown crawler access receives neutral partial credit rather than being reported as a confirmed block.

Older scans

Clearly labelled estimates

Existing reports do not contain bot-specific evidence. They use stored robots/sitemap, metadata, schema, headings, word-count and specificity signals to provide an estimate. A re-scan is required before Clunky AI describes the result as direct.

Interpretation

What the number means

85–100

Strong AI-search foundations

70–84

Visible, with gaps

50–69

Hard to understand clearly

0–49

Hard to retrieve or cite

Important limitations

  • The scan assesses one page, not every page on a domain.
  • It checks whether evidence is visible, not whether it is true.
  • It does not test every CDN, firewall or user-agent-specific response that a real crawler might encounter.
  • It does not guarantee indexing, ranking, recommendation, retrieval or citation by any AI system.
  • Prompt tests in the paid sprint are dated observations and may vary as systems change.

Why these signals?

OpenAI’s publisher guidance says public sites should allow OAI-SearchBot if they want content discoverable in ChatGPT Search. Bing’s AI Performance reporting measures cited pages and grounding queries. Google’s guidance continues to emphasise crawlability, useful original content and ordinary search fundamentals rather than a separate set of guaranteed GEO tactics.

Read the primary guidance from OpenAI, Bing Webmaster Tools and Google Search Central.

Run the free checksView the AI Visibility Sprint