Verified against ChatGPT · 2026-07-30
Check whether ChatGPT and Perplexity can see a specific page right now
A direct test that forces ChatGPT or Perplexity to use their live browsing tool on one exact URL and prove it with a verbatim title/heading quote, instead of answering from stale training data.
The prompt
Ready to copy — highlighted parts are example details you can swap.
Use your live browsing or search tool now — do not answer from memory or training data. If you cannot fetch the URL live, say so plainly instead of guessing. URL TO CHECK: https://example.com/pricing Fetch this exact URL right now and report: 1. Can you access it — yes or no. If no, state the exact error (blocked, 404, timeout, JavaScript-only content with nothing in the initial HTML, and so on). 2. Quote the literal page <title> and the first visible heading, verbatim, in quotation marks. This proves a live fetch — a model without real access cannot produce the exact strings. 3. Check each of these expected facts and confirm present / not found / contradicted: starting price is $49/month, page has a heading called "Simple, transparent pricing", last updated July 2026 4. State today's date as you understand it, so this result isn't confused with a cached or training-data answer from a different point in time. Do not soften a "not found" into "likely" or "probably" — report exactly what you could and couldn't retrieve.
Customize the highlighted detailsoptional — the prompt above already works
Why this works
A chat model without an active browsing/search call will often answer confidently from stale training data rather than admit it can't see the live page — the trap this category most needs to catch. Forcing it to quote the literal <title> and first heading verbatim is a built-in hallucination check: a model that hasn't actually fetched the page cannot reliably produce those exact strings. This also catches failure modes a static homepage crawl can't fully simulate for an in-product search experience — robots.txt blocks on a specific path, JavaScript-only rendering with nothing in the initial HTML, or a noindex tag — on any URL, not just the homepage the automated checker inspects.
What you get back
1. Access: Yes. 2. Title: "Pricing — Example" / First heading: "Simple, transparent pricing" 3. Starting price $49/month: confirmed. Heading text: confirmed, exact match. Last updated July 2026: not found — no visible update date on the page. 4. Today's date (per this fetch): August 2026.
Verified against
ChatGPT GPT-5.1 (Search) · 2026-07-30
Perplexity Sonar Pro · 2026-07-30
Changelog
- 2026-07-30 — Initial publish, built around the verbatim-title-and-heading check after seeing both engines occasionally answer from stale cached knowledge instead of admitting a fetch failed.
Building this for real?
This is a free starting point. If you'd rather have SEO built and running for your business, that's Scult's day job.
EXPLORE SEO
