Sandboxes

Overview

The Koyeb Sandbox SDK provides a Python interface for programmatically creating, managing, and interacting with Koyeb sandboxes — ephemeral compute environments designed for isolated code execution on the Koyeb platform.

Why use sandboxes?

A code sandbox is an isolated, ephemeral environment designed to safely run, test, and experiment with code without affecting other systems or requiring complex setup. It provides developers with a virtualized or containerized execution space where dependencies, environment variables, and runtime contexts can be fully controlled and discarded after use.

You should use a code sandbox to:

  • Execute untrusted or user-generated code securely
  • Prototype applications quickly
  • Test APIs or libraries in clean environments
  • Demonstrate functionality without configuring local infrastructure

Sandboxes are especially valuable in platforms for AI model evaluation, online coding environments, CI/CD pipelines, and educational tools that require safe, reproducible, and on-demand compute environments.

The Koyeb Sandbox SDK

The Koyeb Sandbox SDK includes both synchronous (Sandbox) and asynchronous (AsyncSandbox) classes, allowing developers to:

  • Launch new sandbox instances
  • Wait for readiness and TCP proxy availability
  • Expose or unexpose internal ports
  • Execute background processes and list or terminate them
  • Access sandbox filesystem and execution interfaces
  • Clean up resources automatically

Get started with sandboxes

Sandboxes are available for users on the Starter, Pro, or Scale plans. Learn more about Koyeb subscription plans (opens in a new tab).

Here's how to get started: