Free tool FAQ generator for page text, with schema and HTML
Paste an article or help page. The rules pull out the questions it already asks and the sentences that answer them; click Generate with AI to have a language model write questions in the way readers search. Edit the list, then copy FAQPage JSON-LD and the HTML for the visible FAQ.
- AI questions on request
- JSON-LD and HTML
- Rules run in your browser
Generate an FAQ
The editor follows your text until you edit it. JSON-LD, HTML and the checklist are built from the editor.
Paste the article, Markdown or HTML. At least 80 characters for AI; up to 8,000 are sent.
Language the AI writes in.
One block per question: the question on the first line, the answer below, a blank line between blocks. Same format as the schema markup generator.
AI questions
Written by a language model from your text only. Load them into the editor, then check each answer against the page.
Sample output from 2026-11-17 (qwen3.8-max) for the example text. Change the text and click Generate for your own.
What are UTM parameters?
UTM parameters are tags added to the end of a link so an analytics tool can tell which campaign a visit came from. They do not change the page the link opens.
Which UTM parameters are required?
Google's Campaign URL Builder asks for utm_source, utm_medium and utm_campaign. The parameters utm_term and utm_content are optional and are mostly used for paid keywords and for telling apart two versions of the same ad.
Are UTM parameters case-sensitive?
Yes, Google Analytics 4 and most other tools treat values like Email and email as two different values. You should pick one spelling, usually lowercase, and use it everywhere.
Should I use UTM tags on internal links?
No, because a tagged link between two pages of your own site makes the visit look like it came from that campaign and overwrites the real source. UTM tags should only be used on links that point to your site from somewhere else.
What are utm_term and utm_content used for?
They are optional UTM parameters mostly used for paid keywords and for telling apart two versions of the same ad.
Checklist
- OK4 questions with answers.
- NotePut the same questions and answers visibly on the page (the HTML below). Markup for text readers cannot see is against Google’s structured data guidelines.
- NoteGoogle stopped showing FAQ rich results on 7 May 2026; since August 2023 they had been limited to well-known government and health sites. The markup is still valid schema.org, but it will not change how your result looks in Google.
Preview
How the FAQ looks with the HTML below and no extra styling: native, keyboard-accessible disclosure widgets.
What are UTM parameters?
UTM parameters are tags added to the end of a link, such as utm_source=newsletter, so an analytics tool can tell which campaign a visit came from. They do not change the page the link opens.
Which UTM parameters do I need?
Google's Campaign URL Builder asks for utm_source, utm_medium and utm_campaign. utm_term and utm_content are optional; they are mostly used for paid keywords and for telling apart two versions of the same ad.
Are UTM parameters case-sensitive?
Yes. Google Analytics 4 and most other tools treat Email and email as two different values. Pick one spelling, usually lowercase, and use it everywhere.
Should I put UTM tags on internal links?
No. A tagged link between two pages of your own site makes the visit look like it came from that campaign and overwrites the real source. Use UTM tags only on links that point to your site from somewhere else.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "FAQPage",
"mainEntity": [
{
"@type": "Question",
"name": "What are UTM parameters?",
"acceptedAnswer": {
"@type": "Answer",
"text": "UTM parameters are tags added to the end of a link, such as utm_source=newsletter, so an analytics tool can tell which campaign a visit came from. They do not change the page the link opens."
}
},
{
"@type": "Question",
"name": "Which UTM parameters do I need?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Google's Campaign URL Builder asks for utm_source, utm_medium and utm_campaign. utm_term and utm_content are optional; they are mostly used for paid keywords and for telling apart two versions of the same ad."
}
},
{
"@type": "Question",
"name": "Are UTM parameters case-sensitive?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Yes. Google Analytics 4 and most other tools treat Email and email as two different values. Pick one spelling, usually lowercase, and use it everywhere."
}
},
{
"@type": "Question",
"name": "Should I put UTM tags on internal links?",
"acceptedAnswer": {
"@type": "Answer",
"text": "No. A tagged link between two pages of your own site makes the visit look like it came from that campaign and overwrites the real source. Use UTM tags only on links that point to your site from somewhere else."
}
}
]
}
</script>Paste into the page’s <head> or body. Characters such as < are escaped so the text cannot close the script tag.
<div class="faq">
<details>
<summary>What are UTM parameters?</summary>
<p>UTM parameters are tags added to the end of a link, such as utm_source=newsletter, so an analytics tool can tell which campaign a visit came from. They do not change the page the link opens.</p>
</details>
<details>
<summary>Which UTM parameters do I need?</summary>
<p>Google's Campaign URL Builder asks for utm_source, utm_medium and utm_campaign. utm_term and utm_content are optional; they are mostly used for paid keywords and for telling apart two versions of the same ad.</p>
</details>
<details>
<summary>Are UTM parameters case-sensitive?</summary>
<p>Yes. Google Analytics 4 and most other tools treat Email and email as two different values. Pick one spelling, usually lowercase, and use it everywhere.</p>
</details>
<details>
<summary>Should I put UTM tags on internal links?</summary>
<p>No. A tagged link between two pages of your own site makes the visit look like it came from that campaign and overwrites the real source. Use UTM tags only on links that point to your site from somewhere else.</p>
</details>
</div>The same questions as on-page content. Answers are HTML-escaped.
Version v1.0Updated Update history
What you get and what it is for
The tool gives you three things from one list: questions to check, markup for machines and HTML for people.
AI questions
Phrased the way people search, and they may combine facts from different paragraphs into a new question. The model is told to use only your text, but it can still overstate: in the sample, “Which UTM parameters do I need?” became “…are required?”, which the text does not quite say.
Rule-based questions
Only questions the text already asks, as headings or sentences ending in “?”, with the sentences that follow as the answer. Nothing is invented, but nothing is added either.
FAQPage JSON-LD
Structured data that tells machines which text is a question and which is its answer. It is valid schema.org; Google no longer shows FAQ rich results, so treat it as structure, not as a way to get a bigger result.
Visible HTML
The <details> / <summary> markup readers see. It needs no JavaScript, works with a keyboard and screen readers, and keeps the page and the markup saying the same thing.
A good FAQ answers what readers ask after the main content, in one to three sentences each. It is not a place to repeat the article or to add keywords.
Four steps
You can stay offline the whole time if the rule-based questions are enough.
-
1
Paste the page text
Copy the article or help page. Headings that end in a question mark make the best rule-based result.
-
2
Pick a source
The editor already holds the questions the rules found. Click Generate with AI for questions written by a model; they replace the editor content.
-
3
Edit in the editor
Fix wording, delete weak pairs, add your own. Once you type in the editor, it stops following the text.
-
4
Copy the code
Paste the HTML where the FAQ should appear and the JSON-LD into the same page. Check it with the schema validator.
How questions are found and coded
Rules and code generation run in your browser; only the AI step uses a server.
- Rules: lines that look like headings and end with “?” (or “?”) become questions; the next one to three sentences up to the following heading become the answer, trimmed to about 320 characters (150 for Chinese)
- Questions inside paragraphs (“How long does it take? About three days.”) are picked up the same way; if fewer than two questions are found, ordinary headings are used and flagged for rewriting
- AI: the model is asked for N pairs, phrased as readers search, answers of one to three sentences using only facts stated in the text; missing question marks are added and duplicates removed
- Code: each block in the editor becomes a
Questionwith oneacceptedAnswer;<becomes\u003cinside JSON-LD and answers are HTML-escaped in the markup
| Date | Change | Source |
|---|---|---|
| Until August 2023 | FAQ rich results shown for many sites | Google Search Central documentation |
| 8 August 2023 | Limited to well-known, authoritative government and health sites | Search Central Blog |
| 7 May 2026 | No longer shown in Google Search | Search Central updates page |
| 15 June 2026 | FAQ structured data documentation removed | Search Central updates page |
The example text, rules versus AI
The example text above has four headings that are questions. Here is what each method produced from it on 17 November 2026.
| Question | Rules | AI (qwen3.8-max) |
|---|---|---|
| What are UTM parameters? | Found, answer copied | Same question, answer shortened |
| Which UTM parameters do I need? | Found | Rewritten as “…are required?”: stronger than the text |
| Are UTM parameters case-sensitive? | Found | Same, answer reworded |
| Should I put UTM tags on internal links? | Found | Same, answer reworded |
| What are utm_term and utm_content used for? | Not found (no such heading) | New question built from one sentence |
The rules returned four pairs, all from the page; the AI returned the five asked for, one of them new and one slightly overstated. Both lists are good starting points; neither should be published without a read-through.
FAQ habits that backfire
Most problems come from writing the FAQ for search engines instead of readers.
Markup without visible text
JSON-LD for questions that are not on the page breaks Google’s structured data guidelines. Publish the HTML too.
The same FAQ on every page
Site-wide boilerplate questions add nothing to each page. Write questions about this page’s topic.
Answers the page does not support
AI can produce plausible answers that go beyond your text. Check every claim, especially numbers and “required”.
Expecting rich results
Google no longer shows FAQ rich results. Add an FAQ because readers need it, not for a bigger listing.
See whether readers reach the FAQ
An FAQ only helps if people reach it. TapCub shows which pages keep readers, which lose them and what people searched for before they arrived.
Page reports
Visitors, entries and exit rate per page, so you can see whether a help page keeps readers or loses them.
Learn moreSearch queries per page
Search Console queries show the questions people type before landing on the page: good material for the next FAQ.
Learn moreAI crawlers counted apart
See which AI crawlers read your pages, kept out of the human numbers.
Learn moreFAQ generator questions
Still have a question?
Chat with the team behind TapCub — we usually reply within a few hours on business days.
Does FAQ schema still get rich results in Google?
No. Google limited FAQ rich results to well-known government and health sites in August 2023 and stopped showing them on 7 May 2026. FAQPage markup is still valid schema.org and does no harm, but it no longer changes how your result looks in Google Search.
Then why add FAQPage JSON-LD at all?
It labels which text is a question and which is the answer, which other search engines, tools and crawlers that read schema.org can use. Whether any of them shows it differently is up to them. The visible FAQ is what readers use; the markup is optional.
How many questions should an FAQ have?
As many as readers actually ask, usually three to eight on one page. Each answer should be one to three sentences and link to the full section when there is more to say. The tool lets you pick 3–8 for the AI.
Are the AI answers accurate?
The model is told to use only facts from your text, but it can still reword them more strongly or merge two statements. In the sample above it turned “do I need” into “are required”. Treat the AI output as a draft and check each answer against the page.
What does the extraction without AI do?
It looks for headings or sentences that end with a question mark and takes the following one to three sentences as the answer. It never writes new text, so it works offline and cannot invent facts, but it only finds questions your text already asks.
Is the generated code safe to paste?
Yes. Inside the JSON-LD, < is written as \u003c, so a stray </script> in your text cannot end the script early, and every question and answer in the HTML is escaped. Check the result with the schema validator before publishing.
What happens to the text I paste?
Nothing leaves your browser unless you click Generate with AI. Then the text is sent to TapCub’s API and on to the model provider (currently Alibaba Cloud Model Studio). TapCub does not store it; it keeps only a per-IP request counter and token counts. The limit is 10 AI requests a day, shared with TapCub’s other AI writing tools.
Update history
New features, fixes and data refreshes for this tool, newest first.
-
· v1.0
LaunchedFirst release.
Missing something or found an error? Tell us — requests that come up often go into the next update. All tool updates
Check the markup, then the page
Validate the JSON-LD, make sure the page reads well, and see whether people reach the answers.