Sandbox environment: prices, stock and coupons are synthetic test data. What this means

Can I read any Cartroute page as JSON?

Every public page on cart-route.com (products, categories, retailers, coupons, guides, pricing, status and these docs) returns structured JSON when requested with Accept: application/json or ?format=json. Page views are free, need no key, and are the cheapest way for a browsing agent to read the catalog.

How do I request JSON?

curl -H "Accept: application/json" https://cart-route.com/p/sony-wh-1000xm6
curl "https://cart-route.com/c/tvs?format=json"
curl "https://cart-route.com/docs/errors?format=json"

HTML pages also advertise their JSON form with Link: <…?format=json>; rel="alternate"; type="application/json", and every response carries Link headers pointing to the OpenAPI description and the MCP server card.

Which pages have a JSON form?

PageJSON object
/p/{slug}product with offers, specs and 90-day price_history
/c/{category}category with every product in it
/retailers, /retailers/{slug}retailer_list, retailer
/couponscoupon_list
/search?q=search_result (anonymous preview: 5 a day)
/guides, /guides/{slug}guide_list, guide with live figures
/docs/…doc with the page text
/pricing, /about, /status, /pricing, about, status, page

When should an agent use the API instead?

When it needs filters, sorting, pagination, idempotent retries or more than five searches a day. Page JSON is for reading known URLs, for example a product link a user pasted.