Introducing ComputeSDK: A free and open-source toolkit for running other people's code in your applications

I’m excited to share something we’ve been building that addresses a fundamental challenge in modern application development.

As applications increasingly need to execute user-submitted or dynamically generated code, developers face a critical challenge: running untrusted code is hard and vastly different from running your own code.

The infrastructure complexity problem

I’ve talked with several teams building app builders who have replaced their compute environment multiple times. Each migration required significant engineering effort—rewriting integrations, updating authentication flows, and ensuring feature parity. This pattern of switching providers and rebuilding infrastructure diverts resources from core product development.

This scenario plays out repeatedly across teams building AI agents, educational tools, and any application requiring dynamic code execution. Developers find themselves reaching for something that they thought to be simple, but end up vastly underestimating the complexity involved.

A unified approach to compute

With ComputeSDK we’ve built a unified interface that allows you to hot swap providers at ease. You can think of it like CloudFormation or Terraform, but for running other people’s code. Our simple interface masks considerable technical complexity.

Here’s what implementation looks like:

import { compute } from 'computesdk';
import { e2b } from '@computesdk/e2b';

// Set default provider
compute.setConfig({
  provider: e2b({ apiKey: process.env.E2B_API_KEY })
});

// Create a sandbox
const sandbox = await compute.sandbox.create();

// Execute code
const result = await sandbox.runCode('print("Hello World!")');
console.log(result.stdout); // "Hello World!"

// Clean up
await sandbox.destroy();

Multi-provider benefits

By aggregating multiple providers, ComputeSDK delivers several key advantages:

Vendor neutrality: Like Terraform abstracts cloud infrastructure or CloudFormation standardizes AWS resources, ComputeSDK provides a consistent interface across compute providers. This means your application logic remains portable while you leverage the unique strengths of different platforms.

Operational resilience: When demand spikes or you need additional capacity, you can seamlessly scale across multiple providers rather than being constrained by a single platform’s limitations.

Development velocity: Focus on your application logic rather than learning multiple provider APIs, authentication schemes, and deployment models.

Target applications

ComputeSDK is particularly valuable for:

• AI applications that need to execute LLM-generated code safely
• Educational platforms requiring isolated coding environments
• Developer tools embedding live code execution capabilities
• API services processing user-submitted code securely

The SDK serves teams ranging from individual developers to enterprises, with particular appeal for organizations building agentic workflows where code generation and execution are core components.

Backed by a provider model

ComputeSDK’s architecture is built on top of proven compute providers, each bringing their own strengths to the platform. E2B provides enterprise-grade security through Firecracker microVMs with sub-second startup times. Vercel offers global serverless execution with support for both Node.js and Python runtimes. Daytona delivers full development workspace environments for complex application scenarios.

Rather than reinventing compute infrastructure, we focus on creating the abstraction layer that makes these powerful platforms accessible through a single, consistent interface. For teams requiring additional control or on-premises deployment, we’re also developing premium infrastructure solutions that maintain the same unified API while providing enhanced capabilities. This approach means you benefit from the continued innovation of established providers while maintaining the flexibility to adapt as your needs evolve—whether through cloud providers or self-hosted environments.

Roadmap and availability

ComputeSDK is free and open-source, available now. Near-term development includes:

• A premium Kubernetes-based self-hosted compute environment
• Additional primitives and services (think git, databases, and blob storage)
• Additional provider support (think Neon, Supabase, and/or Turso)
• Enhanced monitoring and observability features

Looking ahead

ComputeSDK represents our bet on the future of development infrastructure: simpler, more flexible, and designed for the AI-assisted development era. We believe developers should focus on creating value rather than managing infrastructure complexity.

The platform is now available on GitHub at github.com/computesdk/computesdk with documentation at computesdk.com/docs. I’d welcome the opportunity to discuss how ComputeSDK might fit into your development workflow and hear about the challenges you’re solving.

Whether you’re building the next generation of AI applications or simply need reliable, flexible compute capabilities, ComputeSDK is designed to simplify that journey.

Garrison Snelling
Founder, ComputeSDK

---

We’re expanding our team. If you’re interested in building developer infrastructure that actually makes developers’ lives easier, let’s connect.