Shopify Catalog, UCP and MCP
Shopify exposes stores to AI agents through three things: Shopify Catalog, the index of eligible products that Shopify shares with AI surfaces such as ChatGPT, Gemini, Copilot and Perplexity; the Universal Commerce Protocol (UCP), the open standard for agents to discover products and complete checkout; and the Storefront MCP server, which lets an assistant search a specific store, build a cart and hand off to checkout. Mojito Labs configures a store for all three and checks the result from the agent's side.
These surfaces change quickly. We track Shopify's changelog and update configurations under a support plan.
The three surfaces, plainly.
What each needs from your store.
Complete product data, honest policies, a checkout without blockers, and measurement.
The same fixes serve all three surfaces and improve Google Shopping, Meta catalogues and on-site search at the same time. The engineering is in the details: attribute coverage by category, availability per market, policy pages that machines can parse.
How we configure it.
Audit, fix, verify from the agent's side, then keep it current.
Verification means asking the assistants the questions your customers ask and reading what they say about your products. The readiness project.
Asked about Catalog, UCP and MCP.
Since June 2026 eligible catalogues are shared by default, with controls in the Shopify admin to adjust or opt out. The point of the audit is to make that a decision you took, not one that happened to you.
No. It adds AI assistants as a channel next to them. The same clean product data feeds all of them from one source.
Yes. The Storefront MCP server gives your own assistant structured access to search, cart and policies. That is an AI systems engineering project with the same approval and measurement discipline as everything else we build.
Quarterly at least. We track Shopify's changelog and editions, and a support plan includes re-verification when a surface changes.


