Free tool

See any site the way ChatGPT sees it.

AI assistants read raw HTML — no JavaScript, no styling. Scan any page to see exactly what they receive: the title, description, structured data, and whether AI crawlers are allowed in at all. No sign-up, no email, and we keep no record of the scan. Run it on a prospect's site and you have the whole first meeting in one screen.

What AI assistants see

Run a scan to fill this report. The values below are what the raw HTML gives an AI crawler — no JavaScript, no styling.

Title: Awaiting scan
Description: Awaiting scan
Main heading (H1): Awaiting scan
Structured data: Awaiting scan

Are AI crawlers allowed in?

GPTBot (ChatGPT)Unknown
ClaudeBot (Claude)Unknown
PerplexityBot (Perplexity)Unknown
Google-Extended (Gemini & AI Overviews)Unknown

JavaScript dependency

Unknown

Low means the content is in the raw HTML. High means a human sees a page that AI crawlers largely do not.

Readable text and page size appear here after a scan.

Reading the result

What each line is telling you

Scanning a prospect's site before the first meeting tends to be more persuasive than any deck. Nothing is stored, so you can run it on anyone.

Title and meta description: This is most of what a model has to work with when it decides whether your client is the right answer. If the title is the CMS default ("Home | Acme") or the description is empty, the page is asking to be skipped. Fixable in minutes.
Structured data: JSON-LD is how a page states facts a model can lift verbatim — what the business does, what it charges, what people ask about it. Zero blocks means the model has to infer everything from prose, and inference is where it invents.
JS dependency: High means the content a human reads is assembled by JavaScript that AI crawlers never run. The page can look perfect in a browser and be nearly empty to GPTBot. This is the one that surprises people.
Crawler access: Whether robots.txt lets each crawler in at all. A single Disallow: / in the wrong group silently undoes every other optimization on the site — and from 15 September 2026, Cloudflare blocks Training and Agent bots by default on ad-bearing pages of newly onboarded domains, which catches GPTBot and ClaudeBot. Worth checking before you blame the content.

Fix what the scan finds — for every client

Preferium optimizes what AI crawlers read and publishes it at the edge — no CMS changes, reversible per element. Resell it under your own brand.

Join the waitlist How white label works