Sandbox MCP is a Model Context Protocol (MCP) server that enables LLMs (MCP hosts/clients) to run code in secure, isolated Docker containers.

While LLMs are really good at generating code, most can't run the code they generate. You end up running this untested code directly on your machine, which could have drastic unintended consequences.

Giving LLMs the tools to test the code safely prevents such issues and helps you generate more accurate code in fewer iterations.

Sandbox MCP gives the LLMs an easy-to-use execution environment that anyone can create and configure through a simple, AI-native MCP server that runs locally.

Inspired by Codapi. Some sandboxes are the same as Codapi sandboxes.

Demo

This demo shows how Sandbox MCP works with Claude Desktop.

Try the video if the GIF isn't clear.

Installation

Download Binary

You can download the appropriate binary for your operating system (for example, Darwin) and processor architecture (arm64) from the Releases page (sandbox-mcp_Darwin_arm64.tar.gz).

Install via Go

Prerequisites:

• Go 1.24 or higher

go install github.com/pottekkat/sandbox-mcp/cmd/sandbox-mcp@latest

Get the path to the sandbox-mcp binary:

which sandbox-mcp

Build from Source

See the Development section.

Usage

Initilization

Before you use sandbox-mcp with LLMs, you need to initialize its configuration:

# Create the configuration directory in
# $XDG_CONFIG_HOME/sandbox-mcp and pull
# the default sandboxes from GitHub
sandbox-mcp --pull

# Build the Docker images for the sandboxes
sandbox-mcp --build

Note

Make sure you have Docker installed and running.

With MCP Hosts/Clients

Add this to your claude_desktop_config.json for Claude Desktop or mcp.json for Cursor IDE:

{
    "mcpServers": {
        "sandbox-mcp": {
            "command": "path/to/sandbox-mcp",
            "args": [
                "--stdio"
            ]
        }
    }
}

Note

Make sure to replace path/to/sandbox-mcp with the actual path to the sandbox-mcp binary.

Available Sandboxes

shell

Run shell commands in a Linux environment with strict security and network constraints.

python

Run Python code with a set of pre-installed libraries.

Important

Your Own Sandbox

You can create and add your own sandboxes in $XDG_CONFIG_HOME/sandbox-mcp/sandboxes. A sandbox is essentially a Dockerfile and a JSON configuration. Check out the examples and the guide to learn more.

network-tools

Use various network tools in an isolated Linux sandbox. The container has network access.

See jonlabelle/docker-network-tools for a list of available tools.

go

Run simple Go code in an isolated sandbox.

apisix

Run Apache APISIX in a sandbox. Custom routes, plugins, and upstreams can be configured through a YAML file and it can be tested using curl commands. Has network access.

javascript

Run JavaScript code using Node.js.

Development

Fork and clone the repository:

git clone https://github.com/username/sandbox-mcp.git

Change into the directory:

cd sandbox-mcp

Install dependencies:

make deps

Build the project:

make build

Update your MCP servers configuration to point to the local build:

{
    "mcpServers": {
        "sandbox-mcp": {
            "command": "/path/to/sandbox-mcp/dist/sandbox-mcp",
            "args": [
                "--stdio"
            ]
        }
    }
}

License

MIT License