The AI Crawler Is Already on Your Store — You Just Can’t See It
Cross-border sellers have spent a decade optimizing for a search engine that shows ten blue links. That era is ending, and most operators I talk to are flying blind into the next one. The discovery layer is shifting to ChatGPT, Claude, Gemini, Perplexity, and AI agents that answer product questions directly — often without a click ever reaching your Shopify or WooCommerce storefront. The uncomfortable part isn’t that AI recommends competitors instead of you. It’s that you have no telemetry on whether AI systems can even read your catalog, which pages they crawl, or what they extract. LLMagnet, a WordPress plugin launched by Ido Navarro and Ben Cohen, is one of the first tools aimed squarely at that blind spot. Whether or not you run WordPress, the problem it names is now every seller’s problem.
What LLMagnet Actually Solves (and What It Doesn’t)
Strip away the launch-day framing and LLMagnet does four concrete things. It tracks real AI bot activity on a WordPress site — which crawlers hit you and which URLs they touched. It produces an “AI Visibility Score,” which the makers describe as a readiness checklist for AI access and understanding, not a measure of whether ChatGPT recommends you. It generates and maintains llms.txt and llms-full.txt files. And it ships an MCP connector that links a WordPress site to AI clients like ChatGPT, Claude, and Cursor.
The distinction the team keeps repeating in the thread matters: crawler activity is tracked separately from recommendation visibility. As Cohen put it, when someone clicks through from an AI platform, that visit can appear in Google Analytics, but LLMagnet also tracks AI crawler requests that happen without any human click. In other words, “views without clicks” means crawler activity, not proof that a shopper saw your brand inside an AI answer. That’s an honest framing, and it’s the kind of nuance most AI-SEO vendors conveniently blur.
Where does the score come from? Navarro broke it down in the comments: crawl frequency, bot quality, traffic volume, page health, and content quality. A thin product or service page with weak headings and missing image alt text drags the content-quality and page-health components down; tighten the structure, recalculate, and the score moves. That’s a defensible mechanism, and it maps to things operators can actually fix.
Why Amazon sellers should care more than Shopify ones
Here’s the counterintuitive take. If you sell primarily on Amazon, you might think AI visibility is a Shopify problem. It isn’t — it’s worse for you. Amazon already owns your customer relationship, your reviews, and your listing copy. When a shopper asks an AI assistant “what’s the best travel espresso maker under $50,” the model synthesizes an answer from whatever it can crawl: review sites, Reddit threads, brand blogs, and increasingly the DTC sites of your competitors. Your Amazon listing is largely invisible to that synthesis. Sellers running a Shopify storefront plus a content layer at least have crawlable assets. Pure Amazon FBA operators have almost nothing the models can quote — unless they’ve built a brand site, and most haven’t.
That asymmetry is why I’d argue the AI visibility gap will hit marketplace-native sellers hardest over the next 24 months. The fix isn’t exotic: a lightweight WordPress or headless content hub, clean structured data, and a way to see what AI crawlers actually ingest. LLMagnet’s WordPress-only focus is a limitation here, but the diagnostic logic transfers.
The incumbents it’s quietly competing with
LLMagnet doesn’t position itself against Ahrefs or Semrush directly, but that’s the real comparison operators will make. Traditional SEO suites measure rankings, backlinks, and organic sessions — all post-click or SERP-based signals. LLMagnet sits earlier in the chain: which AI crawlers hit you, which URLs they reached, how often, and whether your content is machine-readable. The makers also acknowledge that Google is expanding its own reporting, and Cohen conceded they “have to earn our place through the cross-platform picture and useful next steps.” That’s the right posture. Google Search Console will eventually surface more AI-related data; the question is whether a focused third-party tool can deliver actionable recommendations faster than a giant dashboard can.
Where the math breaks
Two honest caveats. First, the score is a readiness proxy, not a revenue metric. A perfect AI Visibility Score doesn’t mean ChatGPT will cite you, and the makers say so explicitly. If you’re a performance marketer who lives in ROAS dashboards, this will feel one step removed from money. Second, WordPress-only is a real constraint. If your storefront runs on Shopify, BigCommerce, or a custom headless stack, LLMagnet can’t touch it today. You can still steal the methodology — audit your robots.txt for AI crawler directives, check whether your key pages have clean headings and alt text, and publish an llms.txt — but you’ll be doing it manually or with a developer.
What Cross-Border Sellers Should Borrow From This Launch
The most valuable thing about LLMagnet isn’t the plugin. It’s the mental model: pre-click visibility is a new funnel stage, and it deserves its own instrumentation. For years, cross-border operators have tracked impressions, clicks, sessions, and conversions. AI discovery inserts a stage before impressions — the moment a model decides whether your content is worth ingesting at all.
Three borrowable moves:
- Audit your AI crawler access. Check your server logs or CDN analytics for user agents tied to OpenAI, Anthropic, Perplexity, and Google’s AI systems. If they’re being blocked by a firewall rule or a lazy
robots.txt, you’re invisible by accident. - Treat
llms.txtlike a sitemap for machines. It’s an emerging convention, not a standard, but publishing one costs nothing and signals intent. Pair it with genuinely clean structured data — product schema, FAQ markup, clear H1s. - Instrument the pre-click layer. Even without LLMagnet, you can tag crawler traffic in your analytics and watch whether the pages AI systems fetch are the ones that actually drive revenue. Cohen’s question in the thread is the right one: are the pages getting crawled the ones that matter most to your business?
The launch thread also surfaced a genuinely useful product direction. A commenter named Daniel Zaitzow pushed the team on whether they’d provide not just a dashboard but specific next steps — “this page needs attention, here’s what’s missing compared with pages being cited.” Cohen’s answer was that the roadmap points toward handing users the proposed change itself, ready to approve. That’s the feature that would separate LLMagnet from every AI-SEO dashboard that just adds another chart. For a cross-border seller juggling listings in three languages and two marketplaces, “here’s the exact edit” beats “here’s your score” every time.
Where My Judgment Says It Falls Short
I like the honesty of the team, and I like that they’re not overselling. But I’d flag four things before an operator commits real budget.
Platform lock-in. WordPress-only excludes the majority of serious DTC operators. WooCommerce stores are a subset; Shopify is the default for cross-border DTC. Until LLMagnet ships a Shopify app or a headless-agnostic crawler, its addressable audience is content bloggers and Woo merchants, not the broader seller economy.
No pricing disclosed. The thread mentions the plugin is free at install, but no tiering, limits, or paid plans are stated in the source. That makes it hard to model as a line item in a tooling stack. Not disclosed is not the same as free forever.
The score’s predictive power is unproven. Navarro’s breakdown — crawl frequency, bot quality, traffic volume, page health, content quality — is reasonable, but nobody in the thread demonstrates that a higher score correlates with more AI citations or more revenue. Until someone runs that correlation on real stores, treat the score as a diagnostic, not a KPI.
Crawler activity ≠ brand visibility. The team is careful about this, but the market will conflate the two. Seeing GPTBot hit your site 400 times a month feels like progress. It isn’t necessarily. It means a machine read you; it doesn’t mean a human was told about you.
The Shopify-shaped hole
If I ran a cross-border Shopify brand, I’d watch LLMagnet closely but not wait for it. The play right now is to replicate the audit manually: pull crawler logs, check AI bot access, clean up your top revenue pages’ structure, and publish an llms.txt. Then, when a Shopify-native equivalent ships — and it will — you’ll already have the baseline data to know whether the tool is moving anything.
What I’d Watch / Test Next
This week, do three things. First, open your CDN or server logs and grep for AI crawler user agents — OpenAI, Anthropic, Perplexity, and Google’s AI systems. If you see zero hits, you have an access problem, not a content problem, and that’s the cheapest fix available. Second, pick your three highest-revenue pages and run a manual readiness audit: clear H1, descriptive alt text, structured data, and an answer to the actual question a shopper would ask an assistant. Third, if you’re on WordPress or WooCommerce, install LLMagnet and compare its crawler log against your own. The gap between what the plugin sees and what your analytics sees is the size of your blind spot — and for most cross-border sellers, that gap is still widening.
