SEO audit API: no-JS vs rendered, diffed

An SEO audit API checks how a page presents itself to search engines and returns the findings as structured JSON instead of a desktop report. This one fetches the URL twice — once as a pure HTTP bot with no JavaScript, once fully rendered in a browser — and returns both views plus the diff, so the gap between them is a field your code can read, not a screenshot you eyeball.

Eventually, usually. Google crawls the raw HTML first and renders JavaScript in a second wave that can lag behind. Content that exists only after JS runs is invisible in that first pass, and other bots — many AI crawlers included — never render at all. That is why comparing the no-JS view against the rendered view is the core of a JavaScript SEO audit.

Raw HTML is what the server sends before any JavaScript executes — what a plain HTTP fetch sees. Rendered HTML is the DOM after scripts have run, which is what users see. On server-rendered pages the two match; on client-side apps the raw view can be nearly empty. The audit quantifies that gap per URL: titles, descriptions, canonicals, h1s and word counts for both views, plus what only exists after rendering.

Search Console's URL Inspection shows Google's own render, but only for sites you have verified and one URL at a time. The audit API approximates both crawler passes on any URL: the no-JS fetch is what a non-rendering bot sees, the rendered fetch is what a rendering crawler sees eventually, and the diff highlights everything that depends on JavaScript.

Yes. The audit works on any public URL — there is no site verification step, because it fetches the page like a bot rather than reading a search engine's private index. Auditing how a competitor's pages render for crawlers is a normal part of technical SEO research.

The diff flags content that only exists after JavaScript runs, titles and descriptions that change between the two views, and canonicals missing from the no-JS pass. Alongside it you get both full views — title, description, canonical, h1, word count each — and the bot-facing metadata: robots directives, Open Graph tags and the JSON-LD types found on the page.

$0.0012 per audited URL — $1.20 per thousand — covering both fetches and the diff. There is no subscription, no seat license and no monthly minimum. A cheaper no-render option returns the no-JS view only. Every account gets $1 of free usage per month, about 830 full audits, with no card required.

Yes — compose it: map lists every URL of the site for $0.0005, then a batch job in summary mode sweeps titles, descriptions, canonicals and status codes at $0.0002 per URL, and you run full audits on the pages that look suspicious. That is a site-wide rendering audit for a few cents, driven from code or CI.

The audit reports the JSON-LD types present and the Open Graph tags as bots receive them, from the no-JS pass — which is the view most crawlers and link-preview scrapers actually use. If your FAQPage schema or og:image only appears after JavaScript runs, that is precisely the kind of gap the report surfaces.

Compare the word counts and h1 between the two views in one audit call. A server-rendered page shows near-identical numbers; a client-side app shows a thin or empty no-JS view against a full rendered one, and the diff lists the JS-only content. It is the fastest programmatic CSR check we know of — one call, one boolean-ish answer.

Both. It is one REST call, so a CI step can fail a deploy when the no-JS view loses its canonical or its content. And the QuanticData MCP server exposes seo_audit as a native tool, so Claude, Cursor or any MCP-capable agent can audit pages and reason over the JSON directly.

Yes. Every account gets $1 of free usage each month with no card required — around 830 audits at list price. The allowance renews monthly and is shared across all QuanticData Data APIs, so the same credit also covers map calls and batch sweeps in a site-wide audit.

No, and that is deliberate. Most audit APIs compress a page into a single number from a fixed checklist — easy to chart, hard to act on, and the weighting is someone else's opinion. This API returns evidence instead: the two fetches, the fields that differ between them, and the bot-facing metadata as it actually is. If you want a score, compute one from the fields with your own weights — your priorities are not ours. What you get here is the raw finding, not a verdict.

Partly, and the honest boundary matters. The no-JS pass is exactly what most AI crawlers get, because the majority do not execute JavaScript — so the diff already tells you how much of your content is invisible to them. The response also reports the bot-facing metadata: robots directives, canonical, Open Graph and the JSON-LD types present. What it does not do is grade your writing for an LLM or score you against an AEO checklist; those are opinions, and this endpoint returns facts.