SDK

Integrate Podstack into your applications using our official Python SDK.

Overview

The Podstack SDK provides a simple, Pythonic interface for managing GPU cloud resources programmatically. Use it to automate deployments, build custom tooling, or integrate Podstack into your ML pipelines.

Quick Example

from podstack import Client

# Initialize client
client = Client()  # Uses PODSTACK_API_TOKEN env var

# Create a GPU pod
pod = client.pods.create(
    name="training-job",
    image="pytorch/pytorch:2.0.0-cuda11.7-cudnn8-runtime",
    gpu_type="A100",
    gpu_count=1
)

# Wait for pod to be ready
pod = client.pods.wait_until_running(pod.id)

# Run training
result = client.pods.exec(pod.id, "python train.py")

# Cleanup
client.pods.delete(pod.id)

Features

• Simple API - Intuitive Python interface
• Full Coverage - Manage pods, VMs, storage, and more
• Type Hints - Full typing support for IDE autocomplete
• Async Support - Optional async/await patterns
• Error Handling - Detailed exceptions for debugging
• Retry Logic - Built-in retry for transient errors

Installation

pip install podstack

In This Section

Coming Soon

• JavaScript/TypeScript SDK
• Go SDK
• Java SDK

Support

For SDK issues, visit Customer Support or check the Troubleshooting guide.