JSON-LD Extractor

How to use

1. 1Paste HTML into the input area. You can paste a full view-source page or just the head block. Click Load sample if you want a worked example first.
2. 2Read the summary for total blocks, valid versus invalid JSON, and the count of distinct Schema.org top-level types.
3. 3Scan the health check for parse errors, missing or wrong @context, missing @type, non-object payloads, and duplicate singleton records like multiple WebSite or Organization blocks.
4. 4Switch the output between Pretty JSON, ready-to-paste Script tags, a Types TSV table, or a Types JSON summary.
5. 5Open each per-block card to see the prettified JSON, then click Copy JSON or Copy script tag to grab a clean version for your CMS, audit doc, or pull request.

About this tool

JSON-LD Extractor parses any HTML you paste, finds every script type="application/ld+json" block, and prettifies the JSON inside. The summary panel shows how many blocks were detected, how many parsed as valid JSON, and how many distinct Schema.org types are declared once @graph members and embedded entities are counted. A health check flags the mistakes that block rich results: malformed JSON (with the parser's error message preserved so you can fix it), missing @context, an @context that does not include https://schema.org, missing @type on the top-level node, top-level values that are not JSON-LD objects, and duplicate singleton records like multiple WebSite, Organization, BreadcrumbList, or WebPage blocks on the same page. Four output views cover every workflow: Pretty JSON shows each block parsed and indented in document order, Script tags wraps the cleaned JSON back into ready-to-paste script tags, Types (TSV) gives you a Type / Count / Blocks table that pastes straight into Google Sheets or Excel, and Types (JSON) is a structured summary for audit pipelines. Each block gets its own card with the detected top-level types, a copy button for the prettified JSON, and a copy button for the full script tag. Parsing uses the browser DOMParser, which builds an inert document, so scripts inside the pasted HTML never run and no remote resource is fetched. Useful for SEO audits, content migrations (verify the new CMS emits the same structured data as the old one), debugging Google Rich Results Test errors, comparing a competitor's structured data setup, validating JSON-LD before pushing it to production, and answering the everyday question: what schema is actually on this page right now? Everything runs locally in your browser, so the HTML you paste here never leaves your device.

Free to use. Works in your browser. No signup, no login.