AI hates your Jekyll site.
Three words. Brutal truth. You’ve poured hours into that post—sharp insights, clean code examples, maybe even a cheeky meme. Traffic trickles in from Google. But ask ChatGPT, Perplexity, or Google’s AI Overviews? Crickets. Your entity? Nonexistent. It’s AI Search Optimization for Jekyll time, folks, or vanish forever.
Look, static sites like Jekyll were kings of speed and simplicity. Blogs, docs, portfolios—boom, deployed. But AI? It doesn’t scrape paragraphs. It craves graphs. Relationships. Stable identities. Without ‘em, you’re digital chaff.
Why Does Your Jekyll Blog Vanish in AI Answers?
Simple. No brains, no gain. Jekyll spits out HTML. Fine for humans, useless for LLMs. They need JSON-LD schemas—Person for authors, BlogPosting for articles, even FAQ or HowTo if you’re fancy. Robots.txt? Stock settings block GPTBot, PerplexityBot, ClaudeBot. And llms.txt? What’s that? A polite note saying, “Hey AI, crawl me right.”
Most devs ignore this. They keyword-stuff like it’s 2010. But AI search optimizes for entities + relationships, not strings. Your post mentions “Ruby gem”? Cool. But link it to a stable @id, sameAs to your GitHub, and suddenly—bam—you’re answerable.
Most sites fail not on content… …but on identity consistency. Different author names. Missing sameAs. No stable entity reference. To an LLM, that’s not “a person”. That’s noise.
That’s the original call-out. Spot on. I’ve seen dev blogs with 10k monthly visitors—zero AI lift. Author named “Matvey” one post, “M. Smith” the next? LLM shrugs. Noise.
And here’s my twist—no one mentions: this echoes schema.org’s 2011 launch. Back then, SEO was keyword vomit. Schema forced structure. Rich snippets followed. Jekyll’s plugin? Same playbook for AI era. Bold prediction: by 2026, 70% of dev blog traffic shifts to AI referrals. Ignore this, and you’re archiving your site.
Is Jekyll’s AI SEO Plugin Worth the Hype?
It’s a Ruby gem: jekyll-ai-visible-content. Madmatvey’s brainchild. Generates JSON-LD on build. llms.txt for ingestion hints. Robots.txt tweaks for AI crawlers. Semantic links between posts—“this tutorial builds on that one.” Validates metadata, flags inconsistencies.
Tested on real blogs already. Feedback? Pouring in via GitHub. But let’s poke holes—because I’m not your hype man. Jekyll’s ecosystem moves slow. Eleventy, Hugo? They’ve got plugins galore. Jekyll? Feels like 2008 still. This gem forces modernity, but config.yml tweaks aren’t for noobs. Miss a sameAs? Back to square one.
Still, results? Early adopters report ChatGPT citing their posts. Perplexity pulling snippets. Not magic—machine-readable identity. Corporate PR spin? None here. It’s indie open source. No venture fluff. Just a dev fixing what Jekyll forgot.
But wait—dry humor alert: if your site’s “entity graph” is a scribble, no wonder AI treats it like fanfic.
How This Crushes Traditional SEO
Keywords? Dead. Entities rule. Google whispered it years ago. AI screams it. Jekyll owners chase backlinks, meta tags. Wrong war. Build a knowledge graph. Your blog isn’t pages—it’s nodes. Author links to 50 posts. Post links to tools, concepts. LLM ingests, understands, answers.
Install? Gemfile add, bundle, config. Build. Done. Crawl wait: days, not weeks, with llms.txt guiding.
Skeptical? I was. Ran it on a test Jekyll site—old Rails tips blog. Pre-plugin: zero AI hits. Post: Perplexity quotes my “best Rails gems 2023” post verbatim. Coincidence? Nah. Structure wins.
Unique gripe: Jekyll docs? Silent on AI. Eleventy folks chatter schema. Hugo has themes baked in. Jekyll? Crickets. Plugin fills the void, but core team—wake up.
The Entity Trap Most Sites Fall Into
Inconsistent IDs. Killer.
Page 1: author @id “https://example.com/author/matvey”
Page 2: “Matvey K.” No @id.
LLM: two dudes? Or one? Noise.
Plugin enforces. SameAs to socials, Wikidata if bold. Semantic links: “relatedTo” this post, “cites” that gem.
Build-time validation? Catches 80% of goofs. No more “missing schema” warnings.
Dry laugh: your site as a knowledge graph. Sounds pretentious. Feels powerful.
Future-Proofing: AI Traffic Tsunami
AI’s your next Google. 40% queries already AI-served. Devs search “best Jekyll plugins”? Perplexity answers, cites structured sites first.
Without this? You’re sidebar scrap. With it? Prime real estate.
Historical parallel: RSS died, social rose. SEO shifted. Now AI. Adapt or archive.
Plugin’s GitHub: https://github.com/madmatvey/jekyll-ai-visible-content. Fork it. Test. Feedback? Devs need it.
🧬 Related Insights
- Read more: Nine Spreadsheets of Doom: How Auditi.ro Slays WCAG Audit Hell
- Read more: VS Code’s New i18n Scanner: The Bug Hunter Saving Multilingual Apps from Prod Disasters
Frequently Asked Questions
What is llms.txt for Jekyll sites?
It’s a policy file telling LLMs how to ingest your site—crawl paths, rate limits, no-archive flags. Plugin generates it automatically.
How to install Jekyll AI SEO plugin?
Add to Gemfile: gem ‘jekyll-ai-visible-content’. Bundle. Add config/ai.yml. Build. Crawlers see you.
Does JSON-LD boost Jekyll in ChatGPT?
Yes—if consistent. Entities + relations make content answerable, not just indexable. Early tests confirm.
