MCPgee
MCPgee
MCPgeeMCPgee

Microsandbox

v1.0.0•Cloud Services•stable

<sup><sup>⭐</sup></sup> - Self-hosted platform for secure execution of AI code. Great for Code Interpreter, Data Analysis, Browser Use.

agentsaicontainerdockerlinux
Share:
6,174
Stars
0
Downloads
0
Weekly
0/5
View on GitHub

What is Microsandbox?

Microsandbox is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to <sup><sup>⭐</sup></sup> - self-hosted platform for secure execution of ai code. great for code interpreter, data analysis, browser use.

<sup><sup>⭐</sup></sup> - Self-hosted platform for secure execution of AI code. Great for Code Interpreter, Data Analysis, Browser Use.

This server falls under the Cloud Services category on MCPgee, the world's largest MCP server directory with 33,000+ servers.

Features

  • <sup><sup>⭐</sup></sup> - Self-hosted platform for secure ex

Use Cases

Secure code execution
Container isolation
Data analysis environment
microsandbox

Maintainer

LicenseApache-2.0
Languagerust
Versionv1.0.0
UpdatedMay 22, 2026
Statushealthy
Maintenanceactive

Works with

ClaudeOpenAIwindowsmacoslinux
View SourceBrowse All Servers

Installation

Manual Installation

npx microsandbox

Configuration

Configuration Details

Config File

claude_desktop_config.json

Performance

Response Metrics

Response Time< 200ms
ThroughputMedium

Resource Usage

Memory UsageLow
CPU UsageLow

How to Set Up and Use Microsandbox

Microsandbox is a self-hosted platform for running AI-generated code inside isolated microVMs, providing a secure environment for code interpretation, data analysis, and browser automation without risking the host system. Built in Rust, it runs on Linux with KVM or macOS with Apple Silicon and exposes a microsandbox-mcp package that AI agents can use via the Model Context Protocol to create sandboxes, execute commands, manage filesystems, and monitor resource usage. It is ideal for giving Claude or other AI agents a safe place to run arbitrary code produced during conversations.

Prerequisites

  • Linux with KVM enabled, or macOS with Apple Silicon (M-series chip)
  • Node.js 18 or later (for the npx MCP installation method) or Rust/Python/Go for SDK usage
  • microsandbox daemon running locally (install via curl or Homebrew)
  • An MCP client such as Claude Desktop or Claude Code
1

Install the microsandbox daemon

Install the microsandbox CLI and daemon on your machine. Choose the method that suits your environment.

# macOS (Homebrew)
brew install superradcompany/tap/microsandbox

# Linux / macOS (curl installer)
curl -fsSL https://install.microsandbox.dev | sh

# npm global
npm i -g microsandbox
2

Start the microsandbox service

Start the local daemon that manages microVM lifecycle. Keep this process running in the background while using the MCP server.

microsandbox start
3

Add the MCP server to Claude

Register the microsandbox-mcp package with Claude Code via the CLI. This uses stdio transport and spawns the MCP server via npx.

claude mcp add --transport stdio microsandbox -- npx -y microsandbox-mcp
4

Configure Claude Desktop manually

If using Claude Desktop, add the following block to your claude_desktop_config.json file.

{
  "mcpServers": {
    "microsandbox": {
      "command": "npx",
      "args": ["-y", "microsandbox-mcp"]
    }
  }
}
5

Verify sandbox creation

Test that the daemon and MCP server are working by creating and running a quick sandbox from the CLI before using it via an AI agent.

# Run a one-shot Python command in a microVM
msb run python -- python3 -c "print('Hello from a microVM!')"

# Create a named persistent sandbox
msb create --name myenv python
msb exec myenv -- python3 -c "import sys; print(sys.version)"

Microsandbox Examples

Client configuration

Claude Desktop configuration using the microsandbox-mcp npm package over stdio transport.

{
  "mcpServers": {
    "microsandbox": {
      "command": "npx",
      "args": ["-y", "microsandbox-mcp"]
    }
  }
}

Prompts to try

Ask Claude to use the sandbox for safe code execution tasks.

- "Create a Python sandbox and run a data analysis script on the attached CSV file"
- "Execute this Node.js snippet safely in a container and show me the output"
- "Set up an isolated environment, install pandas, and generate a histogram from this dataset"
- "Run a security scan script inside a sandbox so it can't affect my host system"

Troubleshooting Microsandbox

microsandbox daemon fails to start on Linux

Confirm KVM is available: run `ls /dev/kvm`. If absent, enable hardware virtualisation in your BIOS or use a hypervisor-compatible cloud instance.

npx microsandbox-mcp fails with connection refused

The microsandbox daemon must be running before the MCP server is started. Run `microsandbox start` (or `msb start`) first, then restart Claude Desktop or Claude Code.

microsandbox is in beta and behaviour is unexpected

Check the latest release notes at https://github.com/superradcompany/microsandbox. As beta software, breaking changes can occur between versions; pin a specific version with `npx microsandbox-mcp@<version>` if stability is needed.

Frequently Asked Questions about Microsandbox

What is Microsandbox?

Microsandbox is a Model Context Protocol (MCP) server that <sup><sup>⭐</sup></sup> - self-hosted platform for secure execution of ai code. great for code interpreter, data analysis, browser use. It connects AI assistants to external tools and data sources through a standardized interface.

How do I install Microsandbox?

Follow the installation instructions on the Microsandbox GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.

Which AI clients work with Microsandbox?

Microsandbox works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.

Is Microsandbox free to use?

Yes, Microsandbox is open source and available under the Apache-2.0 license. You can use it freely in both personal and commercial projects.

Learn More About MCP Servers

Getting Started with MCP

Set up your first MCP server in minutes

MCP Setup Guide

Configure MCP in Claude, Cursor & VS Code

All MCP Tutorials

18+ hands-on guides for developers

MCP FAQ

40+ answers about Model Context Protocol

Microsandbox Alternatives — Similar Cloud Services Servers

Looking for alternatives to Microsandbox? Here are other popular cloud services servers you can use with Claude, Cursor, and VS Code.

Open WebUI

★ 138.2k

User-friendly AI Interface (Supports Ollama, OpenAI API, ...)

Anything LLM

★ 60.4k

The all-in-one AI productivity accelerator. On device and privacy first with no annoying setup or configuration.

LocalAI

★ 46.4k

LocalAI is the open-source AI engine. Run any model - LLMs, vision, voice, image, video - on any hardware. No GPU required.

Nacos

★ 33.0k

an easy-to-use dynamic service discovery, configuration and service management platform for building AI cloud native applications.

Xiaozhi ESP32

★ 26.7k

本项目为xiaozhi-esp32提供后端服务，帮助您快速搭建ESP32设备控制服务器。Backend service for xiaozhi-esp32, helps you quickly build an ESP32 device control server.

Gateway

★ 11.8k

A blazing fast AI Gateway with integrated guardrails. Route to 1,600+ LLMs, 50+ AI Guardrails with 1 fast & friendly API.

Browse More Cloud Services MCP Servers

Explore all cloud services servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.

Cloud ServicesBrowse All Servers

Quick Config Preview

{ "mcpServers": { "microsandbox": { "command": "npx", "args": ["-y", "microsandbox"] } } }

Add this to your claude_desktop_config.json or .cursor/mcp.json

Read the full setup guide →

Ready to use Microsandbox?

Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.

33,000+ ServersFree & Open SourceStep-by-Step Guides
Explore All ServersRead Our Guides