Shopify
Catalogs, variants, prices, and checkout endpoints — covered out of the box.
LiveAgents are commerce's new user.
Paraxor gives your agent the map of any storefront — products, cart, checkout — ready to call. No captchas, no broken scrapers, no per-store glue. Same cart_add works on Shopify, WooCommerce, Magento, or a custom build.
{
"site": "acmehealth.com",
"platform": {
"id": "shopify",
"confidence": 0.97
},
"actions": {
"catalog_query": {
"method": "GET",
"endpoint": "https://acmehealth.com/products.json",
"merchant_label": "Storefront product list",
"side_effects": "reads",
"human_in_loop": "no"
},
"product_detail": {
"method": "GET",
"endpoint": "https://acmehealth.com/products/{handle}.js",
"merchant_label": "Product JSON",
"side_effects": "reads",
"human_in_loop": "no"
},
"inventory_check": {
"method": "GET",
"endpoint": "https://acmehealth.com/products/{handle}/variants.json",
"merchant_label": "Variant stock",
"side_effects": "reads",
"human_in_loop": "no"
},
"pricing_query": {
"method": "GET",
"endpoint": "https://acmehealth.com/products.json?fields=variants.price",
"merchant_label": "Variant pricing",
"side_effects": "reads",
"human_in_loop": "no"
}
}
}Built for the shopping agents reshaping commerce
How Paraxor works
Scraping breaks. Headless browsers get blocked. Per-store integrations don't scale. Paraxor reads any storefront once, classifies what's on it, and exposes the actions through one consistent API your agent already knows how to call.
One request with the storefront URL. We hand back the site already understood — products, cart, checkout, refunds — with no scraping, no headless browser, no captcha to solve.
Shopify, WooCommerce, Magento, BigCommerce, custom builds. Every action gets the same canonical name — so an agent that knows cart_add works against any merchant we cover.
Cacheable on edge, refreshed automatically when a store changes. Your agent spends compute on the user, not on figuring out what each store calls its 'add to cart'.
What your agent gets back
GET /v1/resolve?url=… returns the detected platform and the full manifest of agent-callable actions — each tagged with a canonical name your agent already knows. No HTML parsing, no captcha walls, no per-store glue.
# GET /v1/resolve?url=acmehealth.com
{
"site": "acmehealth.com",
"platform": { "id": "shopify", "confidence": 0.97 },
"actions": {
"catalog_query": {
"method": "GET",
"endpoint": "https://acmehealth.com/products.json",
"side_effects": "reads",
"human_in_loop": "no"
},
"cart_add": { "method": "POST", "endpoint": ".../cart/add.js", /* … */ },
"checkout_initiate": { …, "human_in_loop": "recommended" },
"refund_request": { …, "human_in_loop": "required" }
}
}Coverage
We map every major ecommerce stack so your agent doesn't have to. Start with our coverage, expand on request.
Catalogs, variants, prices, and checkout endpoints — covered out of the box.
LiveREST + REST proxy. Same schema, same automatic refresh cycle.
LiveAdobe Commerce supported via API. Most stores live within minutes.
BetaNative coverage rolling out next. Manual mapping available now.
Coming soonBuilt your own stack? Tell us the URL, we map it to the same schema.
On requestPricing
We're onboarding the first batch of agent platforms by hand. No plan to pick today — usage-based pricing comes later, with a fair window before anything changes for design partners.
while we onboard the first agent platforms
FAQ
We're onboarding the first batch of agent platforms by hand. Tell us what you're building and we'll have you reading real stores within a day.