Shopify Opens Checkout to AI Agents in the Browser

Shopify now lets AI agents running in the buyer's browser complete purchases directly on merchant storefronts.

The company announced WebMCP support for checkout, including Shop Pay, on September 28, 2026. The capability is rolling out to all eligible Shopify merchants, according to the announcement TechCrunch.

The release adds three new WebMCP tools: get_checkout, update_checkout, and complete_checkout. WebMCP is the browser-side connection that lets an agent interact with the storefront. Together the tools let an agent inspect the current checkout state, modify fields such as shipping address or delivery option, and then place the order after buyer authorization.

No screenshots. No scraping. Agents read the live checkout screen through structured calls, apply updates to it, and submit the transaction once the buyer has approved. That authorization gate stays in the browser session, which keeps payment confirmation tied to the shopper rather than to a background process acting on stored credentials.

For developers already working with Shopify's agent stack, the distinction matters. WebMCP is built for agents operating inside the buyer's browser. Shopify's hosted MCP server covers the other path, server-to-server operation where the agent acts from its own infrastructure. MCP, the Model Context Protocol, is a common standard for AI tools to request data and take actions. Both interfaces use Shopify's Universal Commerce Protocol (UCP) as the underlying commerce primitive, the shared set of rules for handling commerce tasks.

Checkout is an extension of WebMCP work Shopify had already shipped for discovery and cart building. Storefronts and carts could previously expose product search and add-to-cart through WebMCP. The developer changelog entry "WebMCP support for checkout," also dated September 28, 2026, states that browser agents can now read and update Shopify checkouts using the new checkout tools Shopify.

That changelog sits alongside two other pieces of the stack. Shopify's Storefront MCP lets AI agents search a store's catalog and manage carts over the Model Context Protocol (MCP) Shopify. Its Global Catalog MCP server extends the same pattern across merchants, enabling agents to search and discover products across the entire Shopify ecosystem. Developer documentation titled "Build commerce agents with UCP" describes agents that authenticate with Shopify and turn carts into checkouts to complete purchases for trusted agents.

Shopify has also provided context for why it is investing here. In Q1 2026, AI-driven traffic to Shopify stores grew 8 times year over year, while orders originating from AI-powered searches increased nearly 13 times Shopify.

The broader context here is a shift from agents that can find products to agents that can transact. Search and cart management made agents useful for research. Checkout makes them accountable for money movement, address handling, shipping selection, and tax and payment state, all areas where ambiguity or stale page parsing creates failed orders and support costs. A structured get, update, and complete sequence reduces that failure surface compared with vision-based automation that guesses at page structure.

In my view, the work for technical teams now moves to edge cases. Buyer authorization needs clear design when the agent proposes a change to delivery speed or address. Merchants will need predictable behavior for discounts, subscriptions, and Shop Pay accelerated flows when the actor is code rather than clicks. Agent builders get structured tools, but they also take on checkout logic that previously lived in frontend validation. That is a reasonable exchange. The long arc of commerce infrastructure has moved toward clearer machine-readable contracts, from APIs to structured checkout, and each step has lowered the cost of building reliable buying experiences.