Can AI find your store?
Run a live scan. 20 seconds, no signup. See whether the engines your buyers ask first can read, trust and quote your store.
Checks the same signals ChatGPT, Claude, Perplexity and Google’s AI use.
Where the search went
Buyers ask an AI before they reach your site.
When a shopper asks ChatGPT or Perplexity for “the best linen shirt under £80”, the answer is assembled from stores the engine can read. If yours is invisible to the crawler, you are not in the running, no matter how good the product is.
- GPTBot reads it
- ClaudeBot reads it
- PerplexityBot reads it
- Google-Extended reads it
Why this matters
Discovery is moving from the search box to the answer.
Roughly 60% of Google searches now end without a click to any website, the searcher gets the answer on the results page itself (SparkToro / Similarweb, zero-click search analysis, 2024). AI answer engines push that further: the model reads a handful of sources it can parse and quotes them directly.
Gartner expects traditional search volume to drop around 25% by 2026 as buyers shift to AI assistants and virtual agents (Gartner, 2024). The stores that get cited are the ones an engine can crawl, parse and trust, structured data, an open crawler policy and a readable feed, not the ones with the most keywords.
This check grades the upstream signals you control. It does not fake-query the engines, it tells you whether they can do their job with your store.
Seven signals, one score
Seven signals decide your score. Every one has a fix.
The scan grades these seven on your live store and rolls them into one 0-100 score. Below is the full map: what each signal is, and the Idukki capability that closes it. In your results, every failing row links straight to its fix.
- 01
AI crawler access
Does robots.txt let GPTBot, ClaudeBot, PerplexityBot, Google-Extended and CCBot in? Block one and that engine cannot read you, so it cannot cite you.
Crawler guidance
We name the exact bots your robots.txt blocks and the one-line allow you need, so every engine can read you.
Fix with Idukki - 02
llms.txt at the root
A short table of contents that points AI agents at your canonical, answer-ready pages. The cheapest single fix, and most stores do not have one.
agentfeed publishes it
Idukki maintains an llms.txt for your catalogue, so AI agents land on your canonical pages instead of guessing.
Fix with Idukki - 03
Structured data (JSON-LD)
Product, Organization and Review schema let an engine read what you sell without scraping prose. Structured data gets cited far more than plain copy.
JSON-LD on UGC + products
Every gallery and tagged product ships Product and Review JSON-LD, the structure engines cite over scraped prose.
Fix with Idukki - 04
Agent-readable product feed
A machine-readable feed of your catalogue (price, availability, image) is what an agent quotes when a shopper asks it to compare or buy.
agentfeed JSON + MCP
A clean product + UGC feed in JSON-LD and over MCP, the format agentic shopping surfaces read directly.
Fix with Idukki - 05
Sitemap + URL structure
A clean sitemap and canonical URLs tell a crawler what exists and which version to trust, so variants do not compete for the same citation.
Canonical-clean surfaces
The scan flags the duplicate and variant URLs splitting your citations and the canonical each should point at.
Fix with Idukki - 06
Quotable metadata
A fact-shaped title and meta description, 120-200 characters, that an engine can lift verbatim into its answer snippet.
Quotable metadata
We show the fact-shaped title and description an engine lifts into its answer, per template, not keyword-stuffed copy.
Fix with Idukki - 07
FAQ / Q&A markup
Questions and answers marked up so an engine can match a shopper question to your answer and quote it back with you as the source.
QnA on your PDPs
Idukki QnA puts marked-up questions and answers on your PDPs, the exact thing an engine matches a query against.
Fix with Idukki
We pass our own test.
idukki.io publishes an llms.txt and llms-full.txt, ships Product and Review JSON-LD sitewide, exposes an agent feed over MCP, and keeps crawler access open. The same fixes we hand you, running on us.
Mouth-to-mouth networking is still, and will keep being, the best marketing resource.
Idukki's unique CMS blends user-generated and brand content, essential for innovative e-commerce strategies.
Questions
The honest answers
- What is llms.txt?
- llms.txt is an emerging open standard, a plain-text file at the root of your site (yourstore.com/llms.txt) that gives AI agents a short table of contents pointing at your canonical, answer-ready pages. It does for AI crawlers roughly what a sitemap does for search engines. Most stores do not have one yet, which makes it the cheapest single win for AI visibility.
- Is it really free, with no signup?
- Yes. The scan runs and shows your score and the first three checks with no account and no card. You only enter an email if you want the full seven-point report and the fixes emailed to you.
- Which AIs does it check for?
- The scan grades the upstream signals that ChatGPT (GPTBot), Claude (ClaudeBot), Perplexity (PerplexityBot) and Google AI (Google-Extended) rely on, plus the Common Crawl bot (CCBot) that several models train on. We grade what you control, your crawler access, structured data, feed and metadata, rather than guessing what any one engine says about you today.
- Will running this change my store?
- No. The scan only reads your public pages, the same way a crawler would. It writes nothing and changes nothing. Any fix it recommends is yours to apply, or Idukki can apply the structured data, feed and llms.txt fixes for you.
- My score is low. Now what?
- A low score means AI engines have little to grab onto, which is fixable and usually fast. Work down the failing checks from the top: crawler access and llms.txt are quick, structured data and a product feed take a little more. Idukki ships all of these, so you can close the gap without a developer sprint.
- Do you store my data?
- The scan itself stores nothing about your store beyond what is needed to return your score. If you ask for the full report, we store the email and URL you give us so we can send it and, if you opt in, re-run the scan weekly. Unsubscribe at any time.
- How is this different from a normal SEO audit?
- SEO audits grade you for the ten blue links. This grades you for the answer. AI engines do not rank pages, they read structured signals and quote a source, so the checks that matter are crawler access, structured data, a readable feed and llms.txt, not keyword density or backlinks.
- Does Idukki pass its own test?
- We run the same checks against idukki.io. We publish an llms.txt and llms-full.txt, ship Product and Review JSON-LD across the site, expose an agent feed over MCP and keep our crawler access open. We would not ship a test we fail.
Ready when you are
Scan it. Then fix it, free.
Run the check on your store, then close every gap with the same platform that powers it. No card to start.
No-code. Live on Shopify, WooCommerce, BigCommerce and more.