🤖 Machine Payable Protocol v1.0 — AI agents can discover, rent, and manage GPUs via HTTP
Base URL
https://gpuhog.com/api
🔑 Authentication
Most endpoints are currently open and do not require authentication. For credit-based rentals and balance checks, include your API key via the X-Api-Key header. You receive an API key automatically when you purchase your first credit pack. Full API key authentication for all endpoints is coming soon.
List Available GPUs
Returns all available GPUs with pricing, tier, and rental info. This is the discovery endpoint for Machine Payable Protocol.
curl example
curl https://gpuhog.com/api/gpus
Response
{
"gpus": [
{
"gpu_id": "NVIDIA GeForce RTX 3090",
"name": "RTX 3090",
"vram": "24 GB",
"price_per_hour_usd": 0.34,
"tier": "piglet",
"rent_url": "https://gpuhog.com/api/rent",
"payment_methods": ["stripe"],
"protocol": "HTTP-402-MPP"
}
],
"protocol": "Machine-Payable-Protocol/1.0",
"docs": "https://gpuhog.com/api/docs"
}
API Docs (JSON)
Returns machine-readable API documentation listing all endpoints.
curl example
curl https://gpuhog.com/api/docs
Response
{
"name": "GPU Hog - Machine Payable GPU Rental",
"version": "2.0",
"protocol": "Machine-Payable-Protocol/1.0 (HTTP 402)",
"base_url": "https://gpuhog.com/api",
"endpoints": [ ]
}
Rent a GPU (Stripe)
Request a GPU rental. Returns HTTP 402 with a Stripe checkout URL for payment. After payment, the pod deploys automatically.
| Parameter | Type | Required | Description |
|---|
gpu_id | string | Yes | GPU RunPod ID or name from /api/gpus |
hours | integer | No | Rental duration (1-720, default 1) |
curl example
curl -X POST https://gpuhog.com/api/rent \
-H "Content-Type: application/json" \
-d '{"gpu_id": "NVIDIA GeForce RTX 4090", "hours": 2}'
Response (402 Payment Required)
{
"status": 402,
"message": "Payment Required",
"protocol": "Machine-Payable-Protocol/1.0",
"order_id": "HOG-a1b2c3d4e5f6",
"gpu": "RTX 4090",
"vram": "24 GB",
"hours": 2,
"total_usd": "0.98",
"payment": {
"method": "stripe_checkout",
"checkout_url": "https://checkout.stripe.com/...",
"expires_at": "2026-03-30T12:30:00+00:00"
}
}
Rent with Credits
Rent a GPU using your token balance. Requires API key. Pod deploys instantly without a Stripe checkout step.
| Parameter | Type | Required | Description |
|---|
gpu_id | string | Yes | GPU RunPod ID or name |
hours | integer | No | Rental duration (1-720, default 1) |
curl example
curl -X POST https://gpuhog.com/api/rent-with-credits \
-H "Content-Type: application/json" \
-H "X-Api-Key: ghk_your_api_key_here" \
-d '{"gpu_id": "RTX 4090", "hours": 3}'
Response
{
"status": "deploying",
"order_id": "HOG-a1b2c3d4e5f6",
"pod_id": "abc123xyz",
"gpu": "RTX 4090",
"credits_spent": 147,
"balance": 853
}
Pig Chat
Chat with the GPU Hog pig assistant. Ask questions about GPUs, pricing, setup, and more.
| Parameter | Type | Required | Description |
|---|
message | string | Yes | Your question (max 500 chars) |
curl example
curl -X POST https://gpuhog.com/api/pig-chat \
-H "Content-Type: application/json" \
-d '{"message": "What GPU should I use for training a 7B model?"}'
Response
{
"reply": "For a 7B parameter model, I would recommend the A100 80GB..."
}
Order Status
Check order status including live pod details, GPU utilization, and SSH/Jupyter access info.
curl example
curl https://gpuhog.com/api/order/HOG-a1b2c3d4e5f6
Response
{
"order_id": "HOG-a1b2c3d4e5f6",
"status": "running",
"gpu": "RTX 4090",
"vram": "24 GB",
"hours": 2,
"pod_id": "abc123xyz",
"access": {
"ssh": "ssh root@203.0.113.42 -p 22042",
"jupyter": "https://gpuhog.com/pod/abc123xyz/"
},
"gpu_stats": [
{ "gpu_util": "23%", "mem_util": "18%" }
]
}
Stop Pod
Stop a running pod to pause billing. Volume data is preserved. Storage costs 5 tokens/hr while stopped.
curl example
curl -X POST https://gpuhog.com/api/order/HOG-a1b2c3d4e5f6/stop
Response
{
"status": "stopped",
"order_id": "HOG-a1b2c3d4e5f6"
}
Start Pod
Restart a stopped or reserved pod. Resumes billing at the GPU hourly rate.
curl example
curl -X POST https://gpuhog.com/api/order/HOG-a1b2c3d4e5f6/start
Response
{
"status": "starting",
"order_id": "HOG-a1b2c3d4e5f6"
}
Terminate Pod
Permanently terminate a pod. All data is deleted. This action cannot be undone.
curl example
curl -X POST https://gpuhog.com/api/order/HOG-a1b2c3d4e5f6/terminate
Response
{
"status": "terminated",
"order_id": "HOG-a1b2c3d4e5f6"
}
Credit Packs
List available token packs with pricing and savings info.
curl example
curl https://gpuhog.com/api/credits/packs
Response
{
"packs": [
{
"id": "pack_500",
"credits": 500,
"price_usd": "5.00",
"label": "Starter Pack",
"savings": ""
}
],
"note": "Credits never expire. 1 token = 1 cent"
}
Credit Balance
Check your token balance and API key. Authenticate with your API key or email.
curl example
curl -H "X-Api-Key: ghk_your_api_key_here" \
https://gpuhog.com/api/credits/balance
Response
{
"email": "you@example.com",
"credits_balance": 1847,
"total_purchased": 2500,
"total_spent": 653,
"api_key": "ghk_your_api_key_here"
}
Machine Payable Protocol
How AI agents interact with GPU Hog
GPU Hog implements HTTP 402 Machine Payable Protocol (MPP), enabling AI agents to autonomously discover, negotiate, and pay for GPU resources:
1
Discover — Agent calls GET /api/gpus to list available GPUs and pricing
2
Request — Agent calls POST /api/rent with desired GPU and hours
3
Pay — GPU Hog returns HTTP 402 with a Stripe checkout URL. Agent or user completes payment.
4
Deploy — Pod spins up automatically. Agent polls GET /api/order/{id} for SSH/Jupyter access.
5
Manage — Agent can stop, start, or terminate pods via API. Or use tokens for instant deploy with POST /api/rent-with-credits.