Quick Start Guide
Deploy your first GPU container on Podstack in just a few minutes. This guide covers the essential steps from account setup to running workload.
Prerequisites
Before you begin:
- Account created and verified
- Wallet topped up with funds
- Project created (or use default project)
Step 1: Add Funds to Your Wallet
- Navigate to Billing > Wallet
- Click Top Up
- Enter the amount (minimum varies)
- Complete payment via Razorpay (UPI, cards, net banking)
- Funds appear in your wallet immediately
Step 2: Create a Project
- Go to Projects in the sidebar
- Click Create Project
- Enter a project name and description
- Click Create
Projects help organize resources and enable team collaboration.
Step 3: Add an SSH Key
- Navigate to SSH Keys
- Click Add SSH Key
- Either:
- Paste your existing public key, or
- Click Generate New Key to create one
- Give the key a name and save
- If generated, download the private key (one-time only)
Step 4: Create Your First Pod
- Go to Compute > Pods
- Click Create Pod
- Configure your pod:
Basic Settings
- Name: Give your pod a descriptive name
- Project: Select your project
- Image: Choose a container image (e.g.,
pytorch/pytorch:latest)
Resources
- GPU Type: Select from available GPUs (A100, H100, V100, etc.)
- GPU Count: Number of GPUs needed
- CPU: Number of CPU cores
- Memory: RAM allocation in GB
Access
- SSH Key: Select your SSH key
- Ports: Configure exposed ports (SSH is typically port 22)
- Review the estimated cost
- Click Create Pod
Step 5: Connect to Your Pod
Once the pod status shows Running:
Via SSH
ssh root@<pod-subdomain>.cloud.podstack.ai
The SSH connection details are displayed on the pod detail page.
Via Web Terminal
Click the Terminal button on the pod card to open a browser-based terminal.
Via Jupyter Notebook
If you enabled notebook access, click the Notebook link to open JupyterLab.
Step 6: Monitor Your Pod
From the Pods list or Pod Detail page, you can:
- View Logs - See container output
- Check Stats - Monitor CPU, memory, GPU usage
- Stop/Start - Pause billing when not in use
- Delete - Remove the pod when done
Cost Management Tips
- Stop pods when not actively using them (billing pauses)
- Right-size resources - don’t over-allocate GPUs or memory
- Use templates - save configurations for quick redeployment
- Monitor spending - check the dashboard for run rates