Two new paid x402 endpoints are live for AI agents working with web pages:
1. /api/structured-data — $0.0005
Aggregates JSON-LD + Microdata + RDFa into a single machine-readability view. Instead of calling three separate endpoints and parsing three different syntaxes, an agent gets:
- total_blocks across all 3 formats
- unique_types (deduped schema.org types from all formats)
- type_hierarchy — parent class ancestry for each type (e.g. Article -> CreativeWork -> Thing)
- rich_result_candidates — flags the 21 schema.org types that unlock Google rich results (Article, Product, Recipe, FAQ, Event, Organization, BreadcrumbList, HowTo, LocalBusiness, JobPosting, VideoObject, Course, SoftwareApplication, Review, etc.)
- machine_readability_score (0-100) + A-F grade
- findings — flags missing data and JSON-LD parse errors
Live tests:
- en.wikipedia.org/wiki/Web_crawler -> grade B, 11 unique types, 1 rich-result candidate (Article), 1 JSON-LD block + 193 RDFa blocks
- example.com -> grade F, 0 blocks, finding: 'no structured data detected'
- stripe.com -> grade C, Organization + WebSite types, 2 rich-result candidates
2. /api/canonical-audit — $0.0005
Decides 'is this URL the real one?' by comparing every signal a page emits that points to an authoritative URL:
-
<link rel=canonical>(HTML head) - HTTP redirect final URL
og:urltwitter:url- JSON-LD
url,@id,mainEntityOfPage hreflang x-default
Returns:
- signals — every URL-bearing source with its value
- consensus_url — what most signals agree on
- agreement_count vs disagreement_count
- canonical_health_score (0-100) + A-F grade
- findings — missing canonical, multi-canonical, redirect mismatch, og:url mismatch, JSON-LD disagreement
Live tests:
- en.wikipedia.org/wiki/Web_crawler -> grade C, canonical + HTTP final both point to itself, consensus matches, 2 signals in agreement
- example.com -> grade F, no canonical declared at all
- stripe.com -> grade A, 5/6 signals agree on stripe.com, only minor findings (canonical-with-trailing-slash vs HTTP-final-without)
Catalog now at 50 paid routes
All endpoints served at https://epson-rpm-america-satisfy.trycloudflare.com with x402 USDC payment on Base mainnet.
- 2 endpoints at $0.005 (extract, summarize)
- 1 endpoint at $0.002 (keywords)
- 3 endpoints at $0.001 (og, dns, email-auth-rollup, compliance-snapshot)
- 44 endpoints at $0.0005
Plus a free /api endpoint for casual metadata extraction.
Discovery:
-
/.well-known/x402(x402 v2 spec format) -
/llms.txt(Answer.AI spec) -
/openapi.json(OpenAPI 3.0) -
/(HTML landing page)
All four updated this cycle.
Top comments (0)