Inventory API reference

The inventory API is public, read-only and needs no key. Use it to build stock alerts, dashboards or procurement tools on top of live CoreGPU availability.

Base URL

https://coregpu.io/api/v1

Responses are JSON, UTF-8, with CORS enabled for every origin. Requests are limited to 60 per minute per IP address.

List GPUs

GET /api/v1/gpus

curl -s https://coregpu.io/api/v1/gpus | jq '.data[] | {slug, available, price_month}'

Example response (truncated):

{
  "object": "list",
  "updated_at": "2026-11-04T15:00:00Z",
  "data": [
    {
      "slug": "h200",
      "name": "NVIDIA H200 SXM",
      "vendor": "NVIDIA",
      "architecture": "Hopper",
      "memory_gb": 141,
      "memory_type": "HBM3e",
      "status": "in_stock",
      "available": 34,
      "allocation_size": 128,
      "max_per_server": 8,
      "price_month": 2699,
      "market_month": 3606,
      "currency": "USD",
      "restock_on": null,
      "url": "https://coregpu.io/gpus/h200"
    }
  ]
}

Get one GPU

GET /api/v1/gpus/{slug}

Returns the same object for a single model, or 404 with {"error":"not_found"}.

Fields

Field	Type	Description
`status`	string	`in_stock`, `low`, `sold_out` or `preorder`
`available`	integer	GPUs free right now in the current allocation
`max_per_server`	integer	Largest server you can order now (1–8, capped by stock)
`price_month`	integer	USD per GPU per month, excluding taxes
`market_month`	integer	Market median reference, USD per GPU per month
`restock_on`	date \| null	Next allocation date (UTC) when sold out or low

Provisioning API (early access)

Endpoints to reserve, renew and stop servers programmatically are in early access. Request access and tell us how you plan to use them.