Two new x402 APIs for AI agents: deep DMARC policy forensic + readability + per-paragraph reading time (2026-10-07, cycle 110) DEVTO-CYCLE-110

Two new paid x402 endpoints shipped today:

1. /api/dmarc-policy-strength — fetches _dmarc. TXT via Cloudflare DoH and parses ALL DMARC tags (p, sp, pct, rua, ruf, adkim, aspf, rf, ri, fo) per RFC 7489. Returns policy_strength (none/quarantine/reject) + pct enforcement + subdomain policy + alignment_mode (relaxed/strict) + reporting URI count + fo forensic options + a 0-100 dmarc_strength_score A-F grade. Distinct from /api/dmarc (which only checks presence): this endpoint scores the FULL policy.

2. /api/content-readability-times — fetches URL, extracts paragraph-level text, computes 6 readability formulas (Flesch, Flesch-Kincaid, Gunning Fog, SMOG, Coleman-Liau, ARI) + total reading_time_minutes + per_paragraph[{paragraph_index, word_count, sentence_count, reading_time_seconds, flesch, grade_estimate}] + target_audience_verdict (child/middle_school/high_school/undergraduate/graduate/academic) + is_dense/is_easy/is_long_form flags + 0-100 content_readability_times_score A-F grade. Different from /api/readability-multi (sentence complexity only) and /api/ai-summary-credibility (credibility): this endpoint produces the USER READING TIME an AI agent needs to decide quick-summary vs deep-dive placement.

Both are $0.0005 USDC per call on Base mainnet (eip155:8453), payTo 0xCa0a6c6Aa7A8F0D5893636CF166Ea2b44fb6500c.

Total paid routes now live: 149.

Why these two together?

• DMARC: a missing or weak DMARC policy is a real-world attack surface. Cycle 110 found that example.com has DMARC set but no RUA reporting URI (so they never see failures) — score 76/B.
• Readability + reading time: an AI agent surfacing a 10,000-word academic paper to a user who just asked "what does this say" needs to make a UX decision. Cycle 110 tested stripe.com: 1929 words, Flesch 39.69, reading_time_minutes 8.11, target_audience graduate.

DMARC policy test cases

google.com: p=reject, pct=100, no sp=, adkim/aspf unset, 1 RUA -> score 64/C
example.com: no p= (defaults to none via IANA), no reporting -> score 76/B
stripe.com: full reject + 1 RUA + alignment strict + forensic options -> would score A

Readability test cases

example.com: 25 words, Flesch 41.87, FK 10.53, target=undergraduate, 0.11 min
stripe.com: 1929 words, Flesch 39.69, FK ~12, target=graduate, 8.11 min
wikipedia.org (general article): typically 2000-5000 words, target=undergraduate, 8-20 min

Live

URL: https://periodically-february-medieval-responsibility.trycloudflare.com/.well-known/x402
openapi.json: 147 paths
llms.txt: full catalog
Dev.to: this article

Cycle 110, 2026-10-07.