llms.txtunread
Fix what AI
actually cites.
Ahrefs analyzed 137,000 sites in May 2026 and found 97% of llms.txt files got zero requests. The plugins that generate them are optimizing something AI doesn't read. Meanwhile every post on your site has fixable structural problems that actually cause citation — buried leads, no question headings, thin entity signals, missing source attribution. This plugin scores every post 0-100 and shows you exactly what to fix.
What AI actually looks at
front_loadedFront-loaded answer
The payload sentence lives in the first ~200 chars, with title terms surfacing early.
question_h2sQuestion-formatted H2/H3
Section headers phrased as questions — AI extracts these as answer targets.
entity_densityEntity clarity
Proper-noun density in the lede — enough anchors for AI to understand who / what / where.
qa_blocksQ&A blocks
Structured Q&A pairs (H2→P, details/summary, bold-question patterns) that AI extracts cleanly.
source_attributionSource attribution
External links to authoritative sources — .gov/.edu bonus, wikipedia/reuters/apnews bonus.
speakable_hintSpeakable coverage
Presence of `speakable` class or SpeakableSpecification-compatible markup.
lede_strengthLede strength
First paragraph is 60-320 chars with 2-4 sentences — not too short, not a wall.
schema_readySchema readiness
Post has excerpt, featured image, and schema-friendly markup patterns.
Free structural scoring. Pro adds AI rubric + rewrites.
Pro is BYOAPIKEY — OpenAI, Anthropic, or Gemini keys stay on your server. Never phone home.
Free
$0 forever · GPL v2
- › Structural scoring on every published post
- › Auto-recalculates on save (no LLM cost)
- › Per-post editor sidebar with score gauge + signal breakdown
- › Dashboard: avg score, histogram, needs-attention top 10
- › Bulk-scan tool for existing catalog
Solo
$79/yr
1 site · everything in Free plus:
- › BYO-LLM rubric scoring — OpenAI / Anthropic / Gemini
- › Inline paragraph rewrites — click "Fix" on weak passages
- › Historical scoring trend
- › CSV export
- › Automatic plugin updates + priority support
Agency
$299/yr
25 sites · everything in Solo plus:
- › 25 site activations — score client catalogs at scale
- › Deactivate + move between clients freely
- › One API key per client (their bill, not yours)
// secure checkout via Stripe · cancel anytime from the license email · 14-day refund on first purchase
You might also need
Same customer base, complementary problems solved. Three-plugin bundles at ~20% off are on the roadmap — join the waitlist.
NewsArticle Schema Pack
Yoast News SEO is $229/yr. This is $79. Same schema, publisher trust, and Google News sitemap.
Schema Conflict Auditor
Find and merge duplicate JSON-LD schema across your entire site.
FishIntel for WordPress
Daily-fresh fishing intel embedded as shortcodes on your charter site.
Questions
Why not just generate an llms.txt file?+
Because AI crawlers ignore them. Ahrefs analyzed 137,000 sites in May 2026 and found 97% of llms.txt files got zero requests — of the requests that did come, 77% were from non-AI bots. Meanwhile every content page has fixable structural problems that actually drive citation: buried leads, no question headings, thin entity signals. This plugin fixes what AI actually reads.
Do I pay for AI on top of the license?+
Yes — but on your own OpenAI / Anthropic / Gemini account, not mine. Pro pays for the plugin (rubric, rewrites, history, bulk, CSV). The LLM calls run against your own key at whatever the provider charges (usually cents per post with cheap models like gpt-4o-mini or claude-haiku).
Do I need Pro at all?+
The free structural scorer catches about 80% of what makes a post citable — front-loaded answers, question headings, entity clarity, source attribution, speakable coverage. Many sites get to a 70+ score just by fixing what free tier flags. Pro adds a full AI rubric with weak-passage callouts and one-click inline rewrites, plus historical trend and bulk analysis.
Will it slow my site?+
No. Structural scoring is pure PHP — no external calls. Runs on save (async). AI rubric only runs when you explicitly click 'Run AI analysis' in the post editor.
Which LLM providers are supported?+
OpenAI (default gpt-4o-mini), Anthropic (default claude-3-5-haiku-latest), and Google Gemini (default gemini-2.0-flash). Model names are configurable. Keys stay on your server — never phone home to kitmobley.