Publish product data compatible tools can read
Active, search-indexable stores publish crawlable product pages, schema.org Product data, a public JSON feed, and an optional MCP server. Compatible tools can use those endpoints when someone connects them. Publishing them does not enroll a store in a partner program, and it does not guarantee traffic.
More ways to be found
Crawlable pages and a public feed give compatible tools something accurate to read when a buyer or operator connects them.
Checkout stays yours
Agents can guide shoppers to the right product, but the buyer still completes payment through your Go!BrightCart checkout.
Ready on every plan
Crawlable pages, structured data, the public JSON feed, and the per-store MCP server are included on every plan, including Launch.
Why This Matters
Your store is prepared for the next buying habit
Shoppers may ask a connected assistant to compare options. Go!BrightCart publishes the product facts those tools need, then hands the buyer to hosted Stripe checkout. Appearance in any assistant is not guaranteed.
Be discoverable sooner
Your active listings expose the product details agents need to understand what you sell.
Reduce buyer effort
Agents can help buyers compare options before sending them to the right checkout path.
Keep the sale controlled
The assistant can start the path, but the buyer completes checkout themselves.
How It Works
From agent search to buyer checkout
Agents can discover, compare, and start a cart — but a human always completes the purchase.
Agent discovers your store
Through the product page, its schema.org data, the public JSON feed, or a connected MCP client. llms.txt is an experimental index in the technical section below, not an enrollment mechanism.
Agent searches and compares
Agent-readable store data lets the assistant answer questions about your products directly.
Buyer completes checkout
A connected tool can start a cart, then hand the buyer to Go!BrightCart checkout. The shopper completes payment. This is buyer handoff, not a partner instant-checkout program.
For the technically curious
Platform APIs and storefront URLs work together
A connected client can use the public feed and MCP server. Product pages stay on each store's preferred canonical URL. llms.txt is an experimental index and is not how a store enrolls with ChatGPT, Claude, or Google.
Platform discovery and actions
Storefront crawl and canonical URLs
Each active store also embeds schema.org Product JSON-LD with offers, price, and availability directly on its preferred product pages: custom domain first, hosted subdomain otherwise.
Common questions
Do I need to do anything to publish agent-readable product data?
Active, search-indexable stores publish crawlable product pages, schema.org Product data, a public JSON product feed, and an optional MCP server. Publishing those endpoints does not enroll the store in ChatGPT, Claude, Google Merchant Center, or any other partner program, and it does not guarantee appearance or traffic.
Can an assistant complete a purchase without the buyer?
No. A connected tool can search the catalog and start a cart, then hand the buyer to Go!BrightCart checkout. The shopper completes payment through the store's hosted Stripe checkout. This is buyer handoff, not a partner instant-checkout program.
Can I opt my store out of search and agent discovery?
Yes. Store-level search-indexing controls cover classic search indexing and the credential-free feed, marketplace discovery, and public MCP catalog. Opted-out stores are left out of those public discovery surfaces. Page-level noindex still applies to individual listings.
Which tools can use this?
Compatible tools can use the published product pages, JSON feed, and MCP server when someone connects them. Go!BrightCart does not automatically submit catalogs to ChatGPT, Claude, or Google. llms.txt is an experimental index, not an enrollment mechanism.
Does this cost extra?
No. Crawlable product pages, the public JSON feed, structured data, and the per-store MCP server are included on every plan, including Launch. Partner feed submission is a separate step and is not turned on by publishing these endpoints.
Publish a storefront that's ready for both kinds of shoppers
Human buyers get a polished storefront. Connected tools get crawlable pages, a public feed, and an MCP server. Same store, no extra plan.