Free SEO tools/AI Content Scorer
Updated 4 Nov 2026AI Content Scorer
See how easily ChatGPT, Google AI Overviews and Perplexity can lift and quote your writing: the opening answer, every section and every sentence, with the exact fix. Enter a page or paste a draft. No signup.
The result
Can AI quote your page?
This free AI content scorer shows how easily ChatGPT, Google AI Overviews, Perplexity and Claude can lift and quote your writing. It checks the opening answer, every section and every sentence, then lists the exact fixes. Enter a web page or paste a draft. No signup.
Every report opens with a plain answer and one coloured tag per check, so you know in two seconds whether anything needs work. Underneath, each section of the page gets its own row: ready to quote, or what to change.
It is an extractability checker, not an AI detector. It does not guess who wrote the text. It counts things you can check yourself: how long the opening is, which paragraphs start with “This” or “It”, how long the sentences are, and whether facts and sources are named.
A real report for one of our own guides, seoshut.com/blog/local-seo-heat-map: easy to quote, with one thing to improve. 21 of its 26 sections are ready as they are.

How to use it
Check your content in five steps
About a minute, from address to fixes.
- 1Choose Web page or Paste text
Score a live page by its address, or paste a draft before you publish it. Pasted text is scored on our server and never stored.
- 2Enter the page or paste the text
For example yoursite.com/blog/your-post. In pasted text, put headings on their own line or start them with #.
- 3Add the question it should answer (optional)
Type the question a customer would ask, for example how long does a roof last. We check whether your opening answers it.
- 4Press Score my content
The report appears in a few seconds, right under the form.
- 5Fix the red rows first
Then the amber ones. Run it again after you edit. Copy the link to share the report, or save it as a one-page PDF.
Your report, explained
What each part of the report means
The top of the same report, numbered.
123456- 1Result
One plain answer: easy to quote, things to improve, or problems to fix.
- 2Checks passed
How many of the checks passed, with one bar segment per check: green passed, amber improve, red fix.
- 3Check tags
One short tag per check, problems first. Hover a tag on a computer to see the numbers behind it.
- 4Fix rows
For every amber or red check, one line that says what to change. A clean page shows none.
- 5Four numbers
Sections ready to quote, reading ease, average sentence length and the words we scored.
- 6Four parts
The opening answer, section by section, sentence by sentence, and facts and sources. Each part shows the text it is talking about.
How it works
How the content is scored
Eleven yes-or-no checks, every rule shown. Checks that do not apply, such as headings on a short text, are left out of the total.
| Check | Passes when | Why it matters |
|---|---|---|
| Opens with a direct answer | The first paragraph under the title is 10 to 90 words, its first sentence is 35 words or fewer, and it does not warm up (“In today’s world”, “Have you ever”) | ChatGPT took 44.2% of its citations from the first 30% of a text (Kevin Indig, 1.2M answers) |
| Answers your question at the top | At least 60% of the key words of your question appear in the title or the first 120 words. If a lower section covers it, you see which one | Optional. Only runs when you type a question |
| Paragraphs make sense on their own | No more than 10% of paragraphs start with a word that points back, such as This, It, They, However or Also | Bing asks for “sentences that make sense even when pulled out of context” (Microsoft Bing) |
| Short, separate paragraphs | No more than 10% of paragraphs are over 120 words | Bing: “Avoid long walls of text” (Microsoft Bing) |
| Sentences are easy to follow | No more than 15% of sentences are over 25 words | GOV.UK’s 25-word guideline (GOV.UK) |
| Clear headings | At least one H2 or H3 heading for every 450 words | Bing: headings “mark where one idea ends and another begins” (Microsoft Bing) |
| Plain, readable English | Flesch reading ease of 50 or more (30 to 49 is improve, under 30 is fix) | The Flesch formula (formula) |
| Specific numbers and dates | At least 1 number, price, percentage or date per 100 words. Never counted as a fix | Adding statistics was one of the best methods in the GEO study |
| Names or links its sources | At least 2 links to other websites (1 for texts under 600 words). Never counted as a fix | Citing sources was one of the best methods in the GEO study |
| Uses lists or tables | At least one list or table in texts of 500 words or more | Bing: lists and tables are “clean, reusable segments” (Microsoft Bing) |
| Shows who wrote it and when | An author name and a published or updated date. Articles only, not pasted text | Google asks for bylines “where one might be expected” (Google) |
A section is ready to quote when its first paragraph names its own subject, is 8 to 120 words long and has no sentence over 35 words. A heading with only sub-headings under it is shown as a group heading and not counted.
Features
What you get in every report
- Every section, in page orderEach H2 and H3 with its first words, its length and a Ready, Improve or Fix status with the reason.
- The sentences to rewrite firstThe longest sentences and the paragraphs that start by pointing back, with the section they are in.
- Your question, checkedType the question a customer asks and see which of its key words your opening covers.
- Score a draft before you publishPaste the text from Google Docs or Word. Nothing is stored.
- Share and saveCopy a link to the result, or save it as a one-page PDF.
- Free, no signupNo email, no card, no trial.
Part 2 of the same report: every section with its status. Section 8 has a 36-word sentence, so it is marked Improve.


The research
What gets quoted, by the numbers
AI answers quote the top of a page most. Kevin Indig studied 1.2 million ChatGPT answers and 18,012 verified citations: 44.2% came from the first 30% of the text, 31.1% from the middle and 24.7% from the end (Search Engine Journal, Feb 2026).
The same study found that cited text was almost twice as likely to use definitive language (36.2% against 20.2%), and that 78.4% of citations containing a question came from headings.
In the GEO study (Princeton and others, KDD 2024, 10,000 queries), citing sources, adding quotations and adding statistics improved visibility in generative engine answers by 30% to 40% on its main measure, while keyword stuffing gave “little to no improvement” (arXiv).
Where ChatGPT’s citations sit in a page
Source: Kevin Indig, 18,012 verified ChatGPT citations, Feb 2026.
Google: no special tricks
“There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary.” A page must be “indexed and eligible to be shown in Google Search with a snippet” (Google Search Central). Clear writing helps people and AI in the same way.
Bing: write snippable answers
Microsoft Bing’s guide asks for “concise, self-contained phrasing in lists, Q&As, and tables” and says assistants “can often lift these pairs word for word into AI-generated responses” (Microsoft Bing, Oct 2025).
Limits
What this tool can’t do
Knowing the limits helps you trust the result you do get.
Use it to
- Find the paragraphs and sentences that are hard to quote, in seconds
- Check a draft before it goes live
- Check that your opening answers the question people ask
- Compare your page with a competitor’s, section by section
- Give a writer a clear, shareable list of fixes
It won’t
- It is not an AI detector. It does not guess whether a person or AI wrote the text. OpenAI withdrew its own AI classifier in July 2023 “due to its low rate of accuracy”; it caught 26% of AI-written text (OpenAI).
- No promise of citations. A clear page is easier to quote, but AI tools also weigh authority, freshness and competition.
- It reads the HTML, not JavaScript. Text that only appears after JavaScript runs is not scored. Paste it instead.
- English rules. The reading ease formula and the “points back” words are built for English.
- It cannot judge truth. It counts numbers and sources; it cannot tell whether they are right.
- Crawling and markup are elsewhere. Robots.txt, AI crawlers and schema are checked by the AEO Grader.
Using AI to write is fine. Google says “Appropriate use of AI or automation is not against our guidelines” and “Using AI doesn’t give content any special gains. It’s just content” (Google Search Central, Feb 2023). What matters is whether the page helps the reader.
Next steps
What to do with your result
Readers skim too. Jakob Nielsen found that on the average web page, people read at most 28% of the words, and 20% is more likely (Nielsen Norman Group). Writing that AI can quote is writing people can scan.
Your next step, by result
Keep it current. Check that AI tools can reach the page, then see whether they actually name you.
Rewrite the sentences in Part 3 first: split anything over 25 words and name the subject where a paragraph starts with “This” or “It”.
Start with the opening. Two or three sentences under the title, about 40 to 60 words, that answer the main question directly.
Add a heading that asks the question in the customer’s words and answer it in the first sentence under it. Find the questions people ask with the prompt miner.
Link every number to the page it came from, and say who published it in the sentence itself.
If the page builds its text with JavaScript, AI crawlers that do not run it see an empty page. Paste the text here, and check what crawlers get.
Want your pages written to be quoted by AI?
Our team rewrites and builds pages for Google and AI answers, for businesses in the US and UK.
FAQs
FAQs
What is an AI content scorer?
An AI content scorer checks how easily AI tools such as ChatGPT, Google AI Overviews and Perplexity can lift and quote a page. This one checks the opening answer, every section and every sentence, facts and sources, and lists the fixes. It does not detect whether AI wrote the text.
How do I measure the extractability of a blog post?
Check whether each part of the post can be quoted on its own. Count how long the opening is, how many paragraphs start with words like “This” or “It”, how many sentences run over 25 words and whether facts and sources are named. This tool counts all of these for you, section by section.
How do I measure sentence-level extractability for AEO?
Look at each sentence on its own. A quotable sentence names its subject, says one thing and is short: GOV.UK suggests breaking up sentences over 25 words. Part 3 of the report lists the longest sentences and the paragraphs that open by pointing back, with the section each one is in.
Is this an AI detector?
No. It does not guess who or what wrote the text. AI detectors are unreliable: OpenAI withdrew its own classifier in July 2023 because of its low accuracy, after it caught only 26% of AI-written text. Google says using AI is fine when the content is helpful.
Is the AI Content Scorer free?
Yes. It is free with no signup. Each visitor can score 40 web pages and 100 pasted texts a day. The same page and question reuse their result for 10 minutes.
Is my pasted text stored?
No. Pasted text is scored on our server and thrown away. We keep only anonymous counts, such as the number of words and which checks passed, never the text itself.
How long should the opening answer be?
About 40 to 60 words is a good target, and the tool flags openings over 90 words. Put the answer in the first sentence, under 25 words if you can. In one study of ChatGPT citations, 44.2% came from the first 30% of the text.
What is a good Flesch reading ease score?
For most websites, 60 or more is plain English and 50 to 59 is fairly difficult. The tool passes 50 or more. Technical pages can score lower and still be fine for their readers, so treat it as a guide, not a rule.
Does a good score mean ChatGPT will quote my page?
No. A good score means the page is easy to quote. Whether an AI tool chooses it also depends on whether it can reach the page, how trusted the site is and what else answers the question. Use the AI Visibility Checker to see what AI actually says.
Does it work for any page?
It works for any public page whose text is in the HTML, in English. If a site blocks our check, or builds its text with JavaScript, paste the text instead. Robots.txt, AI crawlers and schema are checked by the AEO Grader.
Sources
- Google Search CentralAI features and your website (updated Dec 2025)
- Google Search CentralCreating helpful, reliable, people-first content
- Microsoft BingOptimizing your content for inclusion in AI search answers (Oct 2025)
- Search Engine JournalKevin Indig: The science of how AI pays attention (Feb 2026)
- arXivGEO: Generative Engine Optimization (KDD 2024)
- GooglePassage ranking announcement (Oct 2020)
- WikipediaFlesch–Kincaid readability tests (formulas and bands)
- Inside GOV.UKSentence length: why 25 words is our limit (2014)
- Nielsen Norman GroupHow little do users read? (2008)
- OpenAIAI classifier, withdrawn July 2023
- Google Search CentralGoogle Search’s guidance about AI-generated content (Feb 2023)
- Perplexity Help CenterHow Perplexity works (numbered citations)
More free tools
More free AI search tools
Once the writing is ready, these check the rest.
Read the guides: content mapping: give every customer question one page that answers it first, Local SEO heat maps and How to check your keyword rankings.