Is your storefront agentic?

Every agent-readiness check

An agent-ready store lets AI agents reach its pages, read its product data, use its controls and, increasingly, buy through agentic commerce protocols. These are the checks behind the score, grouped by category, each with how we test it and how to fix it.

Access 25% of score

Agents and AI search bots can reach your pages without blocks, CAPTCHAs or walls.

  • Don't block AI shopping agents at the CDN or WAF

    An agent-friendly store serves AI search bots and user-initiated AI agents (ChatGPT-User, Claude-User, Perplexity-User) the same product page a browser gets. WAF rules, bot-protection products and default CDN settings often return a 403, a 402 or a challenge instead, which makes the store invisible to AI shopping answers.

  • robots.txt for AI agents and AI search bots

    An agent-ready robots.txt lets AI search bots (OAI-SearchBot, Claude-SearchBot, PerplexityBot) and user-initiated agents (ChatGPT-User, Claude-User, Perplexity-User) read your homepage and product pages. You can still block AI training crawlers like GPTBot and ClaudeBot without losing visibility in AI shopping answers.

  • CAPTCHAs block AI shopping agents on product and cart pages

    AI shopping agents stop at CAPTCHAs. Google's Gemini computer use model is instructed never to solve or bypass one, and OpenAI's Operator hands control back to the user. An agent-friendly store keeps visible CAPTCHA challenges off the homepage, product pages and cart, and protects only forms like login or contact.

  • Age gates, consent walls and geo-redirects vs AI agents

    Age gates, cookie-consent walls and IP-based geo-redirects can replace your storefront with an interstitial before an AI agent sees a single product. An agent-friendly store keeps the real content in the server HTML, overlays gates client-side, and serves a default region at its homepage instead of forcing a redirect.

Product data 25% of score

Price, stock and variants are in the HTML and in valid Product JSON-LD.

  • Server-render product name, price and stock for AI agents

    AI shopping agents and AI search bots such as ChatGPT, Claude and Perplexity read the HTML your server returns and do not run JavaScript. An agent-ready product page has the product name, the price with its currency, and the stock status in that initial HTML, not loaded later by a client-side script.

  • Product JSON-LD for AI shopping agents

    Product JSON-LD is a schema.org Product block with an Offer (price, currency, availability) in a script tag of the product page's server HTML. AI shopping agents and Google Shopping read it as the machine-readable version of the page; it must include Google's required merchant-listing fields and match the price shoppers see.

  • Model product variants with ProductGroup for AI agents

    Products with sizes, colors or other options should be described as a schema.org ProductGroup with variesBy and hasVariant, where each variant is its own Product with a SKU, price, availability and URL. Without it, AI shopping agents see one product and one price and can't tell whether the size a shopper wants is in stock.

  • Return policy and shipping structured data for AI agents

    AI shopping agents answer "can I return this?" and "when will it arrive?" before recommending a purchase. An agent-ready store publishes both in structured data, a MerchantReturnPolicy and shipping details (OfferShippingDetails or an organization-level ShippingService), on its products or its Organization.

  • OpenGraph product and price tags for AI agents

    OpenGraph meta tags (og:title, og:image, og:url, og:type) plus price tags (product:price:amount and product:price:currency, or og:price:*) give AI agents, chat apps and link previews a lightweight summary of a product page. They're a fallback signal when Product JSON-LD is missing, so this check carries low weight.

Operability 20% of score

Add to cart, variant pickers and search are real, labelled controls.

  • Add-to-cart button for AI agents: a real button in a form

    An agent-ready add-to-cart control is a native button with visible text, such as Add to cart, inside a form that posts to the cart. AI agents find actions through the accessibility tree, where a clickable div has no role, and a plain form post works even before JavaScript loads.

  • Size and color selectors AI agents can use: real form controls

    Agent-friendly variant selectors are native radio buttons or select menus with labels, server-rendered inside the add-to-cart form. Choosing a size or color is where AI shopping agents fail most often on product pages, because custom div swatches expose no role or selected state.

  • Accessible names for AI agents: buttons, links and form labels

    Accessible names are the text labels browsers expose for buttons, links and form fields in the accessibility tree. AI browser agents read that tree to decide what to click, so an icon-only cart button with no name looks identical to the wishlist button. Agent-friendly stores name at least 95% of controls.

  • Site search for AI agents: a GET URL that returns products

    Agent-friendly site search works through a plain GET URL, such as /search?q=socks, that returns a server-rendered results page with product links. AI agents can then search your catalog in one request instead of typing into a JavaScript widget, and learn the URL pattern from a search form or SearchAction markup.

  • Store policy pages AI agents can find: returns, shipping, contact

    Agent-friendly stores link five policy pages from every page, usually the footer: returns or refunds, shipping, terms of service, privacy policy, and contact. AI shopping agents read these before recommending a purchase to answer questions like 'can I return this?', so each link must work and return the policy as server-rendered text.

  • Semantic HTML for AI agents: landmarks, one h1, lang and alt text

    Semantic structure means a product page has one main landmark and a nav, exactly one h1 with the product name, a lang attribute on the html element, and alt text on images. AI agents use these to skip site chrome, confirm they are on the right product, and understand images they can't see.

Discovery 15% of score

Sitemaps, canonicals, fast pages and a machine-readable entry point.

  • XML sitemap for AI agents: list every live product URL

    An agent-ready ecommerce sitemap is an XML file, declared with a Sitemap line in robots.txt, that lists every live product URL with a lastmod date. AI agents and crawlers use it to enumerate your catalog instead of guessing links, so dead or missing product URLs mean products that agents never see.

  • Canonical URLs and soft 404s on product pages for AI agents

    Agent-ready product pages declare their own absolute canonical URL and return a real HTTP 404 or 410 when a product does not exist. A soft 404, where a missing product returns 200 or redirects to the homepage, makes AI agents believe a dead product is live and recommend it to shoppers.

  • Store speed for AI agents: response time and HTML page weight

    AI agents fetch pages under tight time and token budgets, so an agent-friendly store answers in under 800 ms and keeps HTML under 1 MB. Slow server responses and bloated product pages make agents time out, truncate content, or spend their context window on markup instead of product facts.

  • Markdown for AI agents: serve product pages as text/markdown

    A Markdown alternative is a plain-text version of a page served when a client sends Accept: text/markdown, or linked with rel=alternate type=text/markdown. It gives AI agents product name, price and options without HTML noise, using far fewer tokens. It is an emerging, low-weight signal.

  • llms.txt for ecommerce stores: what to include

    llms.txt is a Markdown file at /llms.txt that gives AI agents a curated map of a site. For an ecommerce store it should start with the store name, then link collections, search, shipping, returns, contact and any agent endpoints such as a UCP profile or sitemap. It is a low-weight, contested signal.

Agentic commerce 15% of score

UCP profile, agent endpoints and cart permalinks for checkout by agents.

  • UCP profile at /.well-known/ucp: make your store agent-ready

    A UCP profile is a JSON file at /.well-known/ucp that tells AI shopping agents which commerce capabilities your store supports (catalog, cart, checkout, order) and where to call them. It is the main public discovery file for agentic commerce, used by Google AI Mode and Gemini.

  • Live agent endpoint: MCP or REST shopping API for AI agents

    A live agent endpoint is the MCP or REST API declared in your UCP profile that AI agents call to search products, manage carts and check out. Declaring the endpoint is not enough: it has to answer. For MCP, that means responding to a tools/list request.

  • Cart and checkout permalinks for AI shopping agents

    A checkout permalink is a URL that builds a cart from item IDs and quantities, then redirects the shopper to checkout. It lets any AI agent, even one without protocol support, give a shopper a ready-to-pay link. The UCP permalink capability standardizes the format.

  • OAuth discovery for AI agents: RFC 9728 and RFC 8414 metadata

    OAuth discovery metadata lets an AI agent find out, without custom setup, how to get a shopper's permission to act on their account. Serve /.well-known/oauth-protected-resource (RFC 9728) and /.well-known/oauth-authorization-server (RFC 8414) or openid-configuration.

  • Bonus agent signals: WebMCP, A2A, AI Catalog, API catalog, rel=ucp

    Bonus signals are five emerging agent-discovery standards (WebMCP, A2A agent cards, AI Catalog, RFC 9727 API catalog and the rel="ucp" link) that the scanner credits when found and ignores when missing. They are early-stage or rarely adopted, so a store is never penalized for skipping them.

Run your store on a platform built for agentic commerce

Medusa is open-source, headless commerce with a full Store API agents can call and storefronts you control end to end.

Deploy on Medusa Cloud

Open-source commerce with a full Store API, hosted and scaled for you. Your storefront, your data model.

Deploy on Medusa CloudRead Docs

Storefront skills for coding agents

Teach Claude Code, Cursor or Codex to build storefronts with the patterns these checks look for.

npx skills add medusajs/medusa-agent-skills
View the skill on GitHub