Citation readiness
Build pages an AI can lift
Score how ready a page is for an engine to lift and cite, cover the sub-questions assistants split your topic into, then check Google's live AI surfaces.
AI engines do not read a page the way a person does.
They lift passages, and a page that buries its answer in long unstructured prose gives them nothing to take. Ranking well and being citable are different properties.
- 4sub-scores per page: structured data, crawler access, answer formatting, citability
- 8signals measured from the page, robots.txt and llms.txt before a model reads anything
- 8-12sub-questions the fan-out tool returns for one query
Failed checks name what to change
The AEO scorer judges whether a page is built for an engine to lift: self-contained sections, question-shaped headings, extractable answers. Each failed check names what it judged and what to change.
Help centre: Scoring tools
The sub-questions behind your topic
The fan-out tool shows how an AI search engine splits a broad query into sub-questions. Cover them and the list becomes the outline of the article.
An llms.txt written for you
The generator writes the file telling ChatGPT, Claude and Perplexity what your business does and which pages matter. The free version is the full tool, and the product files the same artifact into your deliverables.
Structured data that parses
The schema validator shows the structured data a page carries and whether it parses. Markup that fails validation is invisible to rich results.
Checked against the live surface
An audit with AI insights on reads Google's AI Overview and AI Mode for your top keywords and lists the cited URLs. Re-audit and the probe reads fresh.
A trend that shows the work landed
The AI Results Tracker keeps presence and citation history per tracked keyword, so a page rewritten for extractability shows up as a citation trend. Passage clarity, schema and direct answers are the levers.
How it works
The sequence, start to finish. Read it in order: each step assumes the one above it has already run.
Score the page
Paste the URL into the AEO scorer and read which checks failed, because a page can rank well and still be un-citable.
Restructure around the fan-out
Give each sub-question a self-contained section with a direct answer an engine could lift, and validate the page's schema while you are there.
Verify with a live probe
Re-run the scorer for before and after, then audit with AI insights on and watch the tracker for the trend.
The other ways people do this
What each approach does, and what happens here instead. Mechanism only, never an outcome claim.
| Approach | How it works | Here |
|---|---|---|
| Writing for the reader alone | How it worksNarrative headings and long prose. It can rank and read well, and it leaves an engine nothing self-contained to lift, because the answer only exists across three paragraphs. | HereThe scorer counts question-shaped headings, lists, tables and paragraph length, then names the checks that failed and what each one judged. |
| A schema plugin | How it worksInjects JSON-LD from templates. It fixes the markup layer and leaves the prose exactly as it was, so a page can carry valid FAQPage schema and still bury its answers. | HereStructured data is one sub-score of four. The validator shows what parses, and the other three read the writing itself. |
| A GEO checklist article | How it worksLists the practices and leaves you to apply them by hand, page by page. It never reads the page you are working on, so the advice is about pages in general. | HereThe URL goes in and the failed checks come back against that page, so the work list is specific enough to hand to whoever edits it. |
| Publishing an llms.txt and stopping | How it worksTells AI crawlers what the business does and which pages matter. It is a directory, not an answer, so a buried answer stays buried. | HereThe generator writes the file, and its presence moves part of one sub-score out of four. The rest of the score is the page. |
Try the job, or learn the subject
The tool below runs here, on this page, without an account. The rest of the free tools and the lessons sit under it.
Schema validator
Validate JSON-LD structured data on a page. It runs here, on this page, with no account.
More free tools, no signup
- Answer engine optimization tool
- Score how ready a page is to be cited by AI search.
- AI query fan-out simulator
- See the sub-questions an AI assistant would ask.
- llms.txt generator
- Build the llms.txt file that tells AI crawlers what your business does.
Lessons in the Academy
- Writing for AI citations5 min · Intermediate
- Passage structure, direct answers and clear claims that make a page easy to quote.
- How AI assistants choose what to cite5 min · Beginner
- Retrieval, passages and entities: the path from a question to the sources an assistant names.
- llms.txt and AI crawlers5 min · Intermediate
- Which AI crawlers exist, what to allow, and what an llms.txt file does and does not do.
Questions people ask
The ones that decide whether this fits, answered with the verdict first.
- Does the scorer change anything on my page?
- No. It fetches the URL, reads the HTML, robots.txt and llms.txt, and returns scores. Nothing is written to your site and no login is involved.
- What does it cost to run?
- The free scorer is free and needs no account. Inside the product, scoring a page is one credit and an audit is one, while re-opening a stored score costs nothing.
- Do I need a developer?
- Usually not. Headings, direct answers and lists are edits in your CMS. Schema and llms.txt are two files, and the generator writes the llms.txt for you.
- What happens when a check cannot be verified?
- It leaves the score rather than failing you. Citability is a model judgement and labelled as one, and when it cannot be produced the overall renormalises over the checks that ran.
- How is AEO different from GEO?
- AEO works on the page. GEO measures the result. Answer engine optimization makes pages extractable through structure, schema and direct answers, while generative engine optimization measures presence across generative results.
- Can a page rank well and still fail the AEO score?
- Yes. Long unstructured prose can rank and still give an AI engine nothing to lift, and the reverse happens too. The audit carries both signals per page.
- Does llms.txt guarantee AI assistants cite me?
- No. It tells AI crawlers what your business does and which pages matter. Citations still depend on whether those pages answer the question in a form an engine can lift.
- Is the free scorer the whole product?
- No. The free tools score one page per run, while the audit carries SEO and AEO signals per page across the whole site, with history kept.
Where this fits
The jobs either side of this one, and the two pages that read it from your side of the desk.
- For agencies: citable pages across a client roster
- For business owners: be the source an answer quotes
Generative engine optimization
Audits probe Google's AI Overview and AI Mode for your top keywords and record the cited URLs. The tracker keeps that as history.
AI article writer
Give it a keyword. It researches, drafts with internal links and images, then scores the result on the same 36 checks that grade audited pages.
Technical SEO audit
One crawl grades your site A+ to F across eight categories and turns every failure into a fix with its URLs attached.
Citation readiness
See it on your own site
A free audit grades your site A+ to F, then shows where this feature pays first. No card required.