Stale by the time it arrives
Daily batch updates make agents recommend products that have changed price or vanished from stock.
Real-time price & inventory infrastructure
The live product API for AI agents.
Get live price, exact variants, stock and shipping back as one answer your agent can act on.
POST /v1/inventory/search● LIVE
intent
“waterproof trail shoes, size 10, under $140”
{
"match": "Trail Runner X",
"variant": "size 10 / moss",
"price": "$119.00",
"landed_cost": "$129.00",
"in_stock": true,
"ships_in": "2 days"
}Target response < 50ms
The hard truth
Shopping data was built for pages. Not agents.
A browser is not an API
Headless scraping burns time and tokens before an agent can answer a simple shopping request.
Intent does not fit a schema
People ask for outcomes. Catalog APIs expect exact categories, filters and SKU-level knowledge.
The discovery layer
One live answer from intent to checkout.
A continuously updated index between the agent and the transaction layer.
Stream
Merchants broadcast price, stock and variant changes as they happen.
Match
Semantic search resolves natural-language intent against the live product graph.
Act
The agent receives landed cost and a checkout-ready handoff over REST or MCP, in the form a Personal Agent Protocol (PAP) session expects.
Built to fit the stack
Consumer AI agent
index.live product API
Checkout / Payments / PAP
< 50ms
Target query latency
Live
Price and stock signals
Zero
Required schema mapping
The team
Built by people who previously worked at Google and Amazon.
Years spent inside large-scale commerce and infrastructure systems. Now spent on the layer agents are missing.
Private preview