Skip to content

Schema tools

One schema workflow. Find the gaps, fix them, keep them fixed.

Most sites run a schema generator once and never touch it again. Baseline Labs gives you the three pieces AI engines actually need: an audit that grades what is already there, a generator that writes the markup you are missing, and a rolling system that keeps it accurate as the site changes.

Run a free schema audit Generate schema free

The generator needs no account. Your first audit is free, no card required.

Three tools, one workflow

Most schema tools do one job: writing a valid block for a page you point them at. That is the middle step. The two around it - knowing what you already have, and keeping it true afterwards - are where structured data actually goes wrong.

Read
Schema Audit
What have I got, and what is broken? Best for diagnosing a site before you change anything on it.
What it does
  • Takes your domain and crawls it
  • Reads JSON-LD, microdata, RDFa, OpenGraph and microformats
  • Returns a graded gap report, page by page
  • Runs on demand, or on a schedule
Free tier
  • Your first audit, through onboarding. No card.
Audit my site
Write
Schema Generator
What should this page be saying? Best for filling the gaps the audit found.
What it does
  • Takes your domain and finds your pages
  • Reads each page and picks the right type from its content
  • Returns copy-paste JSON-LD script tags
  • Runs on demand
Free tier
  • Free. No account, no card.
Generate schema
Maintain
Rolling Schema
How do I stop this going stale? Best for sites that publish or change content regularly.
What it does
  • Takes your domain once, then watches it
  • Re-reads your pages and regenerates only what changed
  • Injects server-side, so crawlers see it without running JavaScript
  • Runs continuously, with a full sweep at least weekly
Free tier
  • 5 pages on 1 domain
Keep it current

Which one do you need?

I do not know what schema I have
Start with the Schema Audit. It crawls your site and reads every structured-data format on it, not just JSON-LD, then grades each page against what Google and the AI engines look for. Generating more markup on top of a site you have not measured just adds to the pile.
I need markup, and I need it now
Use the Schema Generator. Give it your domain, it finds your pages, reads each one, and writes valid JSON-LD you can paste straight into the page. No account, no card, no fields to fill in by hand.
My content changes and the markup drifts
Set up Rolling Schema. Static markup is wrong the moment a price, an opening hour or an article body changes. Rolling Schema re-crawls, regenerates only what actually moved, and injects server-side - so crawlers that do not run JavaScript still see the current version.

How they work together

  1. Audit grades what is on the site today and shows you the gaps, page by page.
  2. Generate writes the markup those gaps need, for free.
  3. Rolling Schema takes over so the fixes do not decay. New products, revised articles and changed prices get re-detected and regenerated without anyone opening a template.

Does schema markup actually help AI visibility?

Structured data does not buy a citation from ChatGPT or Perplexity. What it does is remove ambiguity. Instead of an AI system inferring what your page is about from prose and layout, JSON-LD states it: this is a Product, here is the price, here is the rating. This is an Organization, here is what it does. Google's AI Overviews, Microsoft Copilot and classic rich results all lean on structured data to decide what to surface. Clean schema is a real advantage, and an indirect one. Treat anybody selling it as a citation lever with suspicion.

Why a generator on its own is not enough

Merkle, Rank Math, Schema App, TechnicalSEO.com, Schema Pro and Google's own Structured Data Markup Helper all solve the same single problem: writing or checking one valid JSON-LD block. That leaves two gaps that matter more than the writing step.

You do not know what you are missing

Most sites carry partial, duplicated or invalid schema without knowing it. More markup on top does not fix the markup underneath. Only reading the site does.

AI crawlers do not run JavaScript

Plugin and tag-manager schema is usually injected client-side. GPTBot, ClaudeBot and PerplexityBot never execute it, so that markup is invisible to the exact systems you are trying to reach. Server-side injection is the only guarantee.

Schema audit or schema validator?

A validator - Google's Rich Results Test, or the schema.org validator - checks one page against the spec and returns pass or fail. An audit crawls the whole site, cross-references JSON-LD against the OpenGraph, microdata and RDFa sitting beside it, and shows you patterns: which entity types are incomplete everywhere, which nested nodes have no stable @id, which pages carry nothing at all. A validator is a spot check. An audit is a map.

Schema markup for Irish businesses and .ie sites

Baseline Labs is an Irish company, based in Galway. Both the Schema Generator and the Schema Audit work on any domain, .ie included, with no market-specific limitations and no assumptions about where your business is registered.

That is worth saying out loud because most of the tools in this category are built and documented for the US and UK markets, and an Irish business evaluating them is reading someone else's examples. If you run a .ie site and want a look at what AI engines currently understand about it, the audit is the place to start.

Common questions

There is no single best one, and it depends on whether you need markup written once or kept true. For a page or two, any free generator will do, ours included. For a site that changes, a generator alone is the wrong shape of tool: you need something that re-checks and regenerates, which is what Rolling Schema does.
A generator writes new JSON-LD from your page content. An audit reads what is already published across your site and grades it, so you know what is broken or missing before you write anything new. Run the audit first if you have never measured your schema coverage.
Yes. The Schema Generator needs no account and no card. An account adds saved results, scheduled audits and Rolling Schema's automatic refresh.
Schema does not guarantee a citation. It gives AI systems an unambiguous read of the page rather than forcing them to infer meaning from prose and layout, which makes a page easier to parse correctly and harder to misrepresent. That is a real advantage, but an indirect one.
WordPress plugins write schema at publish time and usually inject it client-side, which AI crawlers cannot read. Rolling Schema re-crawls your domain on a rolling cycle, regenerates only the pages that changed, and injects server-side. It runs on WordPress through our plugin, on Cloudflare through a worker, and anywhere else by pulling the JSON-LD from our API into your own templates.
Whenever the content underneath it changes: a price, an opening hour, an event date, a rewritten article. Hand-maintained schema drifts from that moment until somebody notices. Rolling Schema detects the change and regenerates only the pages it affected.
That depends on your tier and on what the crawl queue is doing, so your Rolling Schema dashboard reports the cadence it is actually running at rather than a number from a brochure. The floor is fixed: every registered domain gets a full sweep at least weekly, whatever its pages say they need.
Yes, and it is the right order to do it in. The Schema Audit reads the JSON-LD, microdata, RDFa, OpenGraph and microformats across your site and shows you the gaps first. Your first audit is free through onboarding.
A tool that writes schema.org structured data, usually as JSON-LD, from a page's own content, so you do not have to hand-write the code or fill in a form field by field.
The markup itself cannot, but wrong markup can. A block with the wrong type, invented ratings, or claims that contradict what is visible on the page can cost you rich-result eligibility or draw a manual action. Accurate schema carries no downside. The risk is bad schema, not schema.
Yes, indirectly. It does not move rankings by itself, but it makes pages eligible for rich results such as star ratings, FAQ dropdowns and breadcrumbs, which lift click-through, and it gives both Google and the AI engines a clearer read of the page.
Look for one that writes stable entity IDs so entities can be cross-referenced between pages, and that fills the recommended properties rather than the bare minimum needed to pass a rich-result test. No generator is enough on its own if the rest of the site has gaps nobody has audited.
Yes. Give the Schema Generator your domain and it discovers your pages from your sitemap, reads each one, detects the content type and writes valid JSON-LD. There are no fields to fill in.
The common ones and then some: Article, Product, FAQPage, BreadcrumbList, Organization, LocalBusiness, Review, Event, Recipe, JobPosting and SoftwareApplication among 20 or more. The generator picks the type from the page's own content rather than asking you to choose.
All three scale by domain rather than by page, so none of them involve pasting URLs one at a time. Start with a Schema Audit for coverage across the whole site, then put Rolling Schema on it: the free tier covers 5 pages on one domain, and a paid plan raises that to 100 pages per domain.
Yes, as a ready-to-paste script tag, free and without an account. Merkle's and TechnicalSEO.com's free tools export JSON-LD too. What none of them do is regenerate it after you have pasted it, which is the whole difference between a generator and Rolling Schema.
Yes. Both the Schema Generator and the Schema Audit work on any domain, .ie included, with no market-specific limitations. Baseline Labs is an Irish company based in Galway, so the support and the guidance are written with the Irish market in view rather than translated from a US playbook.

Start where your site actually is

If you have never measured your structured data, audit it. If you know what is missing, generate it. If it keeps drifting, roll it.

Run a free schema audit
George
Online
0%