Google Turned FAQ Rich Results Off.
Publish It Anyway.
Most FAQ schema generators still show you a screenshot of a Google result with little expandable question dropdowns under it. That result no longer exists. Google narrowed FAQ rich results to well-known government and health sites in August 2023, and then removed the feature outright: its own changelog says the deprecation notice went up on 8 May 2026 because the feature “will no longer appear in Google Search starting May 7, 2026,” and in June 2026 Google deleted the FAQ rich result documentation entirely, “no longer shown in Google Search results.” Nobody gets FAQ dropdowns now — not you, not your competitor, not the hospital.
So why does this tool exist? Because the markup and the
rich result were never the same thing. FAQPage is still a valid
schema.org type, Google's own position is that leaving the markup in place
causes no harm, and its job was always to state — unambiguously, in a format
built for parsers — here is a question our business answers, and here is
the answer.
That shape is exactly what answer engines consume. Large language models, AI search surfaces and retrieval systems are in the business of matching a question to an answer, and structured Q&A hands them a clean one instead of asking them to infer it out of your page layout.
The honest limit: we can't show you a report proving that a specific assistant read a specific FAQPage block. No vendor publishes that, and anyone who tells you they can measure it is guessing at best. What we can tell you is what it costs — a few minutes and a few kilobytes — and that the alternative is leaving your best answers locked inside prose. Publish it for the machines that parse, not for a rich result that is gone.
Your questions and answers
Two rows to start. Add as many as you need — the JSON-LD below updates as you type.
Paste a list of Q&As instead
It understands Q: / A:, Question: / Answer:, any line ending in a question mark followed by its answer, and blocks separated by blank lines. Nothing is imported until you say so.
Page address (optional)
Adds @id and mainEntityOfPage so the block is anchored to a real page. Leave it blank and those keys are simply left out — no placeholder is invented.
Your JSON-LD
Wrapper on: paste straight into your HTML. Wrapper off: raw JSON for a CMS schema field.
Things worth checking
Every warning here is advice, not a gate — it is your content and your call. Nothing you type is sent anywhere: there is no API behind this page, no request is made, and nothing is stored in your browser either. Reload the page and it is a blank slate.
Paste It, Match It,
Check It.
1. Where the block goes. Copy the output
with the script wrapper on and paste it into your page's <head>.
That is the convention, not a requirement — a JSON-LD block inside
<body> is equally valid and gets read the same way. One
FAQPage block per page; don't stack several.
2. The page must actually say it. This is the rule most people break. Every question and answer in your markup has to be visible to a human on that same page. Marking up answers that only exist in the schema is a structured-data guidelines violation — and beyond the guidelines it is simply dishonest, which is the part that eventually costs you. Write the FAQ on the page first, then mark up what you wrote.
3. Answer HTML, kept short. A limited set
of tags is understood inside an answer — h1–h6,
br, ol, ul, li,
a, p, div, b,
strong, i, em. Anything outside that list
gets flagged above rather than silently deleted, because stripping a user's
content without telling them is how generators quietly corrupt work.
4. Check the markup, not the rich result.
Google's Rich Results Test is the wrong instrument here — the feature it reported
on no longer exists, so a silent result from it tells you nothing about whether
your JSON is well-formed. Validate the markup itself at
validator.schema.org,
which still checks FAQPage properly. Then view-source on the live
page and confirm the block actually rendered — plenty of CMSs mangle a pasted
script tag on the way in.
5. The bug we fixed that others didn't.
If an answer of yours contains the characters </script> — say
you are answering a question about code — a naive generator emits them raw, the
browser closes the JSON-LD block early, and the rest of your schema lands on the
page as visible text. This tool escapes it, and tells you it did. Same for the
quotes, backslashes, line breaks and emoji that break string-concatenated
output: everything here is serialised with a real JSON encoder, not glued
together from fragments.
Schema Is One Signal.
Being Readable Is a System.
Perfect FAQ markup does nothing if AI crawlers are blocked before they reach the page — and over half the live business sites we've tested were quietly turning them away. Access, entity schema, content, measurement: we do the whole job, and we test everything instead of assuming it works.