Skip to main content
OpenComputer gives your AI agents a full Linux VM in the cloud. Each sandbox is an isolated Firecracker microVM with its own filesystem, network, and process space — persistent, long-running, and capable of hibernating when idle. Run Claude inside a sandbox and it can write files, install packages, run commands, and build complete projects autonomously.

Claude Agent SDK built in

Claude runs inside the VM with full filesystem and shell access. Think → act → observe loop, streaming events back to your code.

Long-running

Hours or days, not minutes. Install packages, build projects, run test suites, iterate — no cold starts between steps.

Checkpoint & fork

Named snapshots you can fork from. Try five approaches in parallel from the same starting point — like git branches for VMs.

Full Linux VM

Not a container. Real VM with its own kernel, memory, and disk. Hardware-level isolation via Firecracker — the same technology behind AWS Lambda.

​
Install

npm install @opencomputer/sdk

export OPENCOMPUTER_API_KEY=your-api-key
Grab your API key from app.opencomputer.dev

​
Quick Example

import { Sandbox } from "@opencomputer/sdk";

const sandbox = await Sandbox.create();
const session = await sandbox.agent.start({
  prompt: "Create a hello world Express app and test it",
  onEvent: (event) => {
    if (event.type === "assistant") console.log(event.message);
  },
});

await session.done;
await sandbox.kill();

​
Next Steps

Quickstart

Working agent in 2 minutes

How It Works

Architecture: Firecracker VMs, hibernation, checkpoints

Agents

Run Claude inside sandboxes

Sandboxes

The compute primitive