# Lantad: an AI scanner that shows what AI can read on your site

> Lantad is an AI scanner: it fetches your page the way a non-rendering AI crawler does, renders it the way a browser does, and reports the difference. Free, one URL, no signup.

- Canonical page: https://lantad.co
- This file: https://lantad.co/index.md

## Key facts

- **What Lantad is:** An AI scanner and generative engine optimization (GEO) tool.
- **What it answers:** Whether an AI crawler can read this page, and which part of it is lost.
- **Cost:** The scan is free, one URL, no account and no card.
- **Time:** Under a minute for a single page.
- **Composite weights:** prose parity 50, access 25, structure 15, schema 10, out of 100
- **Grade bands:** A at 90, B at 75, C at 60, D at 40, F at 0
- **AI crawler detection:** 15 named crawler tokens read from robots.txt, plus a live server response for every one that publishes a user agent of its own.
- **Scanner user agent:** LantadBot/1.0 (+https://lantad.co/bot)
- **What it never does:** Predict a ranking, guess what a model does with a property, or grade a page it could not measure.

## The four sub-scores behind the AI Visibility Score

| Sub-score | Weight | The question it answers |
| --- | --- | --- |
| Prose parity | 50 | Do your words exist before JavaScript runs? |
| Access | 25 | Are the named AI crawlers allowed to fetch this page? |
| Structure | 15 | Can a machine find the answer inside the page? |
| Schema | 10 | Does the JSON-LD say who published it? |

## What one scan does, in order

| Step | What happens | Why it is the only way to see the gap |
| --- | --- | --- |
| 1. Fetch as a crawler | GET the page as LantadBot/1.0 (+https://lantad.co/bot), no JavaScript | This is the copy a non-rendering AI crawler receives |
| 2. Render as a browser | Full render, scrolled and settled | This is the copy a human receives |
| 3. Compare | Shingled text comparison of the two | The difference IS the invisible part of your site |
| 4. Detect crawlers | robots.txt evaluated per token for 15 crawlers, plus a live probe where a user agent is published | Robots rules and real server behaviour can disagree |
| 5. Grade and rank | Composite score, then fixes ordered by what they recover | A finding with no next step is trivia |

## What an AI scanner is, and why this one fetches twice

An AI scanner checks what an AI crawler can actually take from a page, which is not what a browser shows you. Most AI crawlers do not run JavaScript. A site can look perfect on screen and hand ChatGPT, Claude or Perplexity an almost empty shell.

That is why every scan fetches twice: once as a non-rendering crawler and once in a real browser. PROSE PARITY is the gap between them, and it is the sub-score most tools cannot produce at all, because a tool that only reads rendered output has already lost the comparison. It carries 50 of the composite, the largest single weight, because a page whose words only exist after hydration is close to invisible to the crawlers that build AI answers.

## AI crawler detection: every named crawler, one scan

Lantad reads your robots.txt and reports the rule that applies to each of the 15 AI crawler tokens in its registry, then sends a live request as every crawler that publishes a user agent of its own and reports what your server actually did.

Both halves are necessary because they disagree. A robots.txt can allow a crawler while a firewall, a CDN rule or a bot-management setting blocks it, and only the live request finds that. Three of the 15 tokens publish no user agent at all, so they can be read from robots.txt and never probed: the report says which, rather than implying a probe that never happened.

The tokens are grouped by purpose, because blocking a training crawler and blocking a search crawler are different decisions with different costs. A training fetch affects future models. A search fetch is what makes a page citable in an answer today.

## Generative engine optimization, and where the score comes from

Generative engine optimization (GEO) is the work of being readable, and therefore citable, by AI answer engines. It overlaps with SEO and is not the same job: a page can rank well and still hand a crawler nothing.

The AI Visibility Score is a single A to F grade built from four sub-scores (prose parity 50, access 25, structure 15, schema 10, out of 100), with the evidence behind every point. The weights are published rather than hidden, and they are read from one constant in the codebase, so this page and the scorer cannot disagree. Grade bands: A at 90, B at 75, C at 60, D at 40, F at 0.

Alongside it, Entity Confidence is a separate diagnostic over your structured data, reported as a tier rather than a grade. It never feeds the composite, because a checklist score and a continuous composite are different measurements and putting them on one scale would make both harder to read.

## Why this matters now

More people ask an assistant instead of scrolling a results page. Those answers are built from what AI crawlers could read on your site, not from what your visitors see. If a crawler gets a blank page, the answer about you is assembled without you.

This is also why the honest framing is readable rather than recommendable. Being readable is a precondition. Nobody outside those systems can observe what a model does next, and any tool claiming otherwise is guessing.

## Who it is for

Site owners who want to know whether an assistant can describe them correctly. In-house teams who own AI visibility and need evidence rather than opinion. Agencies running the same check across many client sites.

Sites drift: a redeploy, a CDN rule, a theme update. AI readability falls silently, and every AI answer about you is wrong until somebody notices. Monitoring is the part that notices, weekly, on every page you track.

## What it refuses to do

It does not predict rankings, and it does not claim to know what any model does with a property, because nothing outside those systems can observe that.

It does not emit a grade for a page it could not measure. There are seven scan outcomes and only one of them is a grade; the rest state what was read and what was not. A confident wrong grade is worse than no grade, and this is the rule the whole product is built around.

## Common questions

**What is an AI scanner?**

An AI scanner checks what an AI crawler can actually take from a page, which is not what a browser shows you. Most AI crawlers do not run JavaScript, so a site can look complete on screen and hand ChatGPT or Perplexity an almost empty shell. Lantad fetches your page as a non-rendering crawler, renders the same URL in a real browser, compares the two, and grades the difference. Free, under a minute.

**What is generative engine optimization (GEO)?**

GEO is the work of being readable and therefore citable by AI answer engines rather than only by search engines. It overlaps with SEO and is not the same job: Google renders JavaScript and most AI crawlers do not, so a page can rank well and still hand a crawler nothing.

**What is AI crawler detection, and how does it work?**

It is finding out which AI crawlers can reach your pages, and it has two halves because they disagree. One reads robots.txt and resolves the rule for each named crawler token. The other sends a live request as that crawler and records what your server did, which catches a CDN or bot-management rule blocking a crawler robots.txt allows. Lantad checks 15 tokens and probes every one that publishes a user agent. Three publish none, so they can be read and never probed, and the report says which.

**Is optimising for AI the same as Google SEO?**

No. Google renders JavaScript and most AI crawlers do not, so a site can rank well on Google and still be invisible to ChatGPT or Perplexity.

## Related

- https://lantad.co/methodology (markdown: https://lantad.co/methodology.md)
- https://lantad.co/bot (markdown: https://lantad.co/bot.md)
- https://lantad.co/research (markdown: https://lantad.co/research.md)
- https://lantad.co/pricing (markdown: https://lantad.co/pricing.md)
- https://lantad.co/tools (markdown: https://lantad.co/tools.md)
- https://lantad.co/tools/ai-crawlers (markdown: https://lantad.co/tools/ai-crawlers.md)
- https://lantad.co/glossary/geo (markdown: https://lantad.co/glossary/geo.md)

---

Lantad measures whether AI crawlers can actually read a page: it fetches as a non-rendering
crawler, renders as a browser, and reports the gap. Free scan, one URL, no signup.

Method and weights: https://lantad.co/methodology | All pages as markdown: https://lantad.co/md | Crawler policy: https://lantad.co/bot