Shopify and Vercel announced a joint effort to create ecommerce sites that can serve both human shoppers and artificial-intelligence agents, unveiling new development tools at Vercel’s Ship 26 conference in San Francisco on Oct. 15.
Conference reveals new partnership
The two leaders emphasized that the collaboration is intended to simplify the creation of custom storefronts while exposing product data in a format that external agents can read and act upon.
Both executives noted that merchants will be able to reach new types of traffic that rely on machine-readable signals, and that developers will gain a clearer path to integrate AI-driven experiences without rebuilding core commerce logic.
Hydrogen SDK rebuilt for flexibility
By separating concerns, the toolkit reduces the amount of boilerplate required for typical storefront features.
According to the release, the toolkit provides direct access to Shopify data for cart handling, customer account linking, and checkout routing, removing the need to build these components from scratch. It also includes type-safe utilities that help keep data consistent across the front end and any downstream services.
Developers who adopt the new SDK can choose to work with React, Solid, or any other supported library, giving them the freedom to align the technology stack with existing skill sets or performance goals.
Vercel Shop template adds agent capabilities
The template ships with structured data, automatically generated sitemaps, and an LLMs.txt file that guide AI agents through the site.
Product pages can return Markdown that includes descriptions, prices, options and inventory levels, enabling agents to retrieve information without parsing HTML. This design choice means that an external model can request a single endpoint and receive all necessary commerce details in a clean, machine-readable format.
Compared with earlier chatbot integrations, this approach extends the storefront’s data layer rather than overlaying a separate conversational interface, offering a more seamless experience for both humans and machines.
AI shopping assistant and merchant controls
Vercel Shop includes an optional AI-powered assistant that can answer shopper queries, suggest items, adjust cart contents and provide policy details via chat, using the same cart infrastructure as the visual storefront. The assistant draws on the same product catalog that powers the public site, ensuring consistency between what a human sees and what the model reports.
The assistant cannot complete purchases; customers must finish checkout through Shopify’s standard flow, which continues to verify prices, availability and permissible cart modifications. This safeguard preserves the existing security and fraud-prevention mechanisms that merchants rely on.
Developers must enable the assistant manually and configure access to Vercel’s AI Gateway, with options to tailor guidance and presentation of products. Settings allow merchants to restrict which categories are exposed, define response tone, and limit the depth of information returned to the model.
Vercel’s own merchandise store serves as a live demonstration of the template and its built-in shopping assistant, showing how the combined solution works in a real-world environment.
Impact on the development community
Early adopters can clone the repository, run the example locally, and see how the LLMs.txt file influences agent behavior.
Community members have already begun experimenting with custom prompts that pull inventory counts directly from Shopify, allowing a virtual assistant to inform a shopper that a specific size is out of stock before the user even reaches the product page.
