SDK Overview

Official SDK libraries for smol machines. Create and manage microVM machines for secure code execution. The engine runs in-process — there is no server to start.

Available SDKs

The package is published as smolmachines on both npm and PyPI. In Python you import it from the smol module.

Features

• Machine Management - Create, start, stop, delete machines
• Command Execution - Run commands directly in the microVM
• Container Support - Run OCI containers (Docker images) inside machines
• Volume Mounts - Mount host directories into machines
• Resource Control - Configure CPU and memory limits
• GPU Acceleration - Vulkan access to host GPU for rendering and ML
• Local or Cloud - Same Machine API on the embedded engine or the smolfleet cloud

Requirements

The engine runs in-process — there is no server to start. The native (local) path runs on:

• macOS with Apple Silicon (M1/M2/M3), or
• Linux with KVM support (/dev/kvm accessible)

The cloud transport works anywhere the package installs. See Installation for details.

Quick Example

import { Machine } from 'smolmachines';

const machine = await Machine.create({ name: 'my-machine' });
try {
  const result = await machine.exec(['echo', 'Hello!']);
  console.log(result.stdout);
} finally {
  await machine.delete();
}

from smol import Machine, MachineConfig

# Context-managed: the machine is stopped and deleted automatically on exit.
with Machine.create(MachineConfig(name="my-machine")) as machine:
    result = machine.exec(["echo", "Hello!"])
    print(result.stdout)

Next Steps

• Installation - Install the SDK
• Quick Start - Run your first machine
• Machine API - Full API reference
• GPU Acceleration - Enable GPU for rendering and ML