GPUy.dev
RunPod logo

The Complete RunPod Onboarding & GPU Rental Guide (2026)

Beginner

Estimated setup time: 3 minutes•Updated for 2026 infrastructure

Start on RunPod
KYC Policy
None (Zero Mandatory KYC)
Payment Methods
Credit / Debit Card • Crypto (USDC / USDT / BTC) • Prepaid Balance
Min. Deposit
$10.00
Billing Model
Per-second billing with prepaid wallet balance
Trial Credits
None (Promo codes apply on deposit)
Terminal Access
Direct Web Terminal & One-Click Public SSH Key

What You Need Before Starting

✓A valid email address or GitHub/Google account
✓An SSH Public Key (e.g. id_ed25519.pub) for terminal access
✓$10 balance via credit card or crypto for account activation

Step-by-Step Setup Walkthrough

6 Steps to First Run
1

Account Registration & Anti-Fraud Whitelisting

Step 1 of 6

Sign up via Google, GitHub, or direct email without ID verification delays.

  • •Navigate to the official RunPod registration page through GPUy's direct partner link.
  • •Register with your developer email or connect your GitHub profile for streamlined key sync.
  • •No passport or ID selfie is required; accounts are activated instantly upon email verification.
Pro Tip: Use the same email as your GitHub account to sync your public SSH keys automatically.
2

Add Billing Balance (Card or Crypto)

Step 2 of 6

Load your prepaid balance to unlock Secure and Community Cloud clusters.

  • •Go to Billing in the left navigation menu.
  • •Choose 'Add Funds' — credit cards (Visa, MasterCard, Amex) and cryptocurrencies (USDT, USDC via CoinGate) are fully supported.
  • •The minimum initial deposit is $10.00. Funds never expire and are consumed per-second while compute pods are active.
Pro Tip: For developers outside the US without international cards, USDT / USDC on Polygon or TRC-20 carries virtually zero transaction fees.
3

Generate and Add Your SSH Public Key

Step 3 of 6

Enable passwordless terminal access to any rented GPU pod with an ed25519 key.

  • •Open PowerShell (Windows) or Terminal (macOS/Linux) and generate an ed25519 key pair.
  • •Copy the contents of the public key file and paste it into RunPod Settings → 'SSH Public Keys'.
Generate SSH Key & View Public Key
bash
ssh-keygen -t ed25519 -C "runpod-gpu"
cat ~/.ssh/id_ed25519.pub
4

Deploy Your First GPU Pod (On-Demand vs Spot)

Step 4 of 6

Select your desired GPU architecture, container template, and storage volume.

  • •Navigate to 'Pods' → 'Deploy'. Choose between 'Community Cloud' (cheaper consumer cards like RTX 4090) or 'Secure Cloud' (Tier-3 datacenters with H100/A100).
  • •Select your Docker template: RunPod PyTorch 2.4, Fast Stable Diffusion, or vLLM.
  • •Allocate Container Disk (ephemeral local NVMe) and Network Volume (persistent across pod restarts).
  • •Choose 'On-Demand' for uninterrupted training or 'Spot' for up to 60% discount on non-critical jobs.
Critical Warning: Spot pods can be interrupted with a 30-second warning if demand surges. Never run un-checkpointed training on Spot without a Network Volume.
5

Connect via SSH or Web JupyterLab

Step 5 of 6

Access your running container within 15 seconds of launch.

  • •Click 'Connect' on your active pod in the dashboard.
  • •Option A: Click 'Connect to Web Terminal' or 'Connect to HTTP Service [Port 8888]' for instant web UI access.
  • •Option B: Copy the provided SSH command directly into your local terminal.
Direct SSH Connection Example
bash
ssh [email protected] -p 22222 -i ~/.ssh/id_ed25519
6

Stopping vs Terminating: Prevent Runaway Storage Billing

Step 6 of 6

Crucial cost-saving rule: Stopping a pod still incurs container disk retention charges.

  • •Clicking 'Stop' frees the GPU compute cost, but Container Disk and Network Volume continue to bill at ~$0.10/GB/month.
  • •To completely stop ALL charges, transfer your model checkpoints to an S3/Hugging Face repository and click 'Terminate' (trash icon).
  • •Always set up a spending limit alert in Account Settings to prevent surprise depletion.
Critical Warning: Leaving a stopped pod with 200GB container storage will cost ~$20/month even if the GPU is never powered on. Always Terminate unneeded instances!

Insider Cost-Saving & Anti-Waste Tactics

#1: Use Spot pods for inference benchmarks and hyperparameter search to save 50-65% on compute rates.
#2: Store heavy datasets in external S3/Cloudflare R2 buckets instead of large persistent RunPod volumes.
#3: Use Community Cloud RTX 4090 ($0.34/hr) for fine-tuning LoRAs instead of expensive H100 ($2.49/hr).

Frequently Asked Questions (RunPod Onboarding)

Q:Does RunPod require ID verification or KYC documents?

No. RunPod does not require passport, national ID, or selfie verification for standard accounts. You only need a verified email and a valid payment method (credit card or crypto).

Q:Can I pay for RunPod with USDT or cryptocurrency?

Yes. RunPod accepts cryptocurrency payments via CoinGate. You can deposit USDT, USDC, Bitcoin, and Ethereum directly into your balance.

Q:What is the difference between Community Cloud and Secure Cloud on RunPod?

Secure Cloud instances run in Tier 3/4 SOC2-certified enterprise datacenters with redundant power and business-grade bandwidth. Community Cloud instances are hosted by decentralized third-party providers and hobbyists, offering lower prices (e.g. RTX 4090 at $0.34/hr) but varying thermal thresholds.

Q:How do I avoid getting charged when I'm not using RunPod?

To stop all charges, you must click 'Terminate' (not just 'Stop'). A 'Stopped' pod holds your disk and incurs storage charges (~$0.10/GB/month).

Official Developer Onboarding Channel

Ready to Launch on RunPod?

Followed the guide? Jump directly to the official console using our verified partner route with guaranteed zero markup and active referral tracking.

Compare Live RatesGo to RunPod Console