Documentation

Everything you need to get started with GPU NOW.

Getting Started

1. Sign In

Click “Sign In” in the navigation bar to authenticate with Google. Then connect your Solana wallet (Phantom, Solflare) to make payments.

2. Browse Available GPUs

Head to the Marketplace to see all currently available hardware. We only show GPUs with confirmed stock — what you see is what you can rent right now.

3. Rent a GPU

Click the rent button on any GPU. You'll be prompted to approve a SOL transaction (1 hour deposit). Once confirmed, your instance deploys in under 60 seconds.

4. Connect to Your Instance

Go to your Dashboard to see connection details. You'll get SSH access, TCP port mappings, and Jupyter Lab URL once the instance is ready.

Pricing

How Pricing Works

Each GPU has an hourly rate in USD. When you rent, we convert the 1-hour cost to SOL using the live market price. You pay a 1-hour deposit upfront.

  • Prices are per-hour, billed to your account
  • SOL conversion uses real-time market price
  • Payment goes directly on-chain — fully transparent
  • Terminate anytime from your Dashboard

Example Pricing

GPUUSD/hr~SOL/hr
RTX 3080$0.17~0.001 SOL
RTX 4090$0.34~0.002 SOL
RTX 6000 Ada$0.74~0.005 SOL
H100 SXM$2.69~0.017 SOL

* SOL amounts vary with market price. Exact amount shown at checkout.

Connecting to Your Instance

SSH Access

Once your instance is running, you'll see an SSH command in your Dashboard:

ssh root@213.173.102.12 -p 29615 -i ~/.ssh/id_ed25519

Configure your SSH key in the Dashboard settings before connecting.

Pre-installed Software

  • CUDA 12.4
  • PyTorch 2.5
  • Python 3.10
  • Ubuntu 22.04
  • Jupyter Lab
  • Full root access

FAQ

What if a GPU goes out of stock?

We deploy the instance first, then request payment. If stock changes, you get a clear error and no payment is taken.

What if I reject the transaction?

The instance is automatically terminated. You won't be charged.

How do I stop charges?

Go to Dashboard and click “Terminate”. Charges stop immediately.