Can AI agents read your Shopify store?

We don't promise rankings — nobody honestly can. Sichta verifies that your store meets the technical requirements agents need to read, understand and transact with it. What agents do with legible data is up to them; illegible data guarantees you're not in the conversation.

The ten questions a buyer typically asks an agent before purchasing — shipping, returns, stock, materials, sizing and so on. Sichta checks each sampled product against all ten and reports how many it can answer, per product and on average — a direct measure of whether an agent could complete the sale.

Plain-text and Markdown indexes that tell AI systems what your store offers — products, collections, policies. Sichta checks that yours resolve, are not thin, and match your live catalog — and generates a richer version free, auto-maintained on Starter and above.

No. Scanning is read-only. The optional fixes (alt text, structured-data embed) are applied through Shopify's own APIs, can be reviewed first, and can be undone for 30 days. Publishing a generated llms.txt or agents.md means pasting it into your own theme file.

Yes — every Shopify plan. Shopify enables UCP on stores by default, so there is nothing to install for it. Password-protected storefronts are marked blocked, not scored, until you save the password under Settings → Preview mode.

Sichta reads product and shop data only, stores no customer data, and deletes everything on uninstall. Your app data is hosted in the EU (Frankfurt); opt-in AI snapshots send only your store name, domain and a generic question to US-based AI models under Standard Contractual Clauses — never customer data.

Yes. Sichta ships a Sidekick app extension: ask Sidekick “can AI agents read my store?” or “what is failing in my readiness scan?” and it answers from Sichta's latest scan — score, failing checks and fix guidance. Running a new scan stays in the app.

Headless or custom storefronts get partial value: the theme app-embed and Shopify's native /llms.txt, /agents.md and /.well-known/ucp all belong to the Online Store front end. The scanner itself is not Liquid-specific — it fetches your storefront over HTTP and parses the HTML — but if there is no Online Store front end to fetch, most of what Sichta checks has nothing to attach to.

Because "served" is not the same as "correct." Shopify's defaults are generic — they don't reflect your actual catalog, policies or product data. Sichta checks that your llms.txt and agents.md resolve, aren't thin placeholders, and match what you actually sell; runs a 26-check readiness score across structured data, catalog completeness and buyer policies; and generates a richer version free, keeping it current as your catalog changes on Starter and above.