Run AI.
Without the risk.

Isolated virtual machines for AI agent workloads. Not containers. Not shared runtimes. Each sandbox gets its own kernel.

sandbox-7f3a

$ sandbox.create()

Booting microVM...

✓ Ready in 247ms

$ sandbox.commands.run("python3 agent.py")

Running in isolated kernel...

{"status": "complete", "output": "..."}

$ sandbox.kill()

✓ Destroyed. No state remains.

$ _

Powering code playgrounds for

learningrust.orglearningtypescript.orglearningjavascript.orglearningsql.orglearningphp.orglearningzig.org

Built for performance

Fast enough to forget
it's a VM.

0ms

Boot time

0

Playgrounds

0h

Free tier / month

0

Shared state

Developer experience

Five lines of code.

$ npm install @omnirun/sdk

$ cat run.ts

import { Sandbox } from '@omnirun/sdk'

const sandbox = await Sandbox.create('playground')
const result = await sandbox.commands.run('echo "Hello from a microVM"' 							)
console.log(result.stdout)
await sandbox.kill()

$ npx tsx run.ts
Hello from a microVM
✓ Sandbox killed

How isolation works

Three steps.
No complexity.

01

Spin up a sandbox

One API call. Typical boot in ~250ms. Full Linux environment with its own kernel.

02

Run anything

Execute code, install packages, read and write files. Nothing shared between sandboxes.

03

Tear it down

Destroyed completely. No state leaks. No cleanup. Gone.

Use cases

Built for what you're building.

Capabilities

Everything you need.
Nothing you don't.

Not a container.
A virtual machine.

Each sandbox runs its own Linux kernel inside a microVM, the same technology behind AWS Lambda. Hardware-level isolation, not process boundaries.

microVMIsolatedOwn kernel

Network blocked by default

Outbound connections blocked unless you add specific destinations to an allow list.

Files require a signed key

Upload and download URLs expire. Access tokens scoped to a single sandbox.

Start building
in 5 minutes.

Free tier includes 25 sandbox-hours per month. No credit card required.

Get Free API Key