MCP

🚀 HOPX MCP Server

The Model Context Protocol (MCP) server for HOPX. It enables AI assistants to execute code in isolated cloud containers.

MCP Name: io.github.hopx-ai/hopx-mcp

🚀 Quick Start

The HOPX MCP Server allows AI assistants to execute code in isolated cloud containers. Here's how you can get started.

📦 Installation

uvx hopx-mcp

Get your API key at hopx.ai.

✨ Features

• Fast Sandbox Creation: Takes about ~200ms.
• Rapid Container Startup: Only ~0.1s.
• Auto-cleanup: Configurable, default is 600s.
• Internet Access: Enabled by default.
• Complete Isolation: For each execution.
• JWT-based Authentication: Ensures security.

📚 Documentation

Configuration

Get API Key

Sign up at hopx.ai.

Configure IDE

Add MCP server configuration to your IDE:

{
  "mcpServers": {
    "hopx-sandbox": {
      "command": "uvx",
      "args": ["hopx-mcp"],
      "env": {
        "HOPX_API_KEY": "your-api-key-here"
      }
    }
  }
}

{
  "mcpServers": {
    "hopx-sandbox": {
      "command": "uvx",
      "args": ["hopx-mcp"],
      "env": {
        "HOPX_API_KEY": "your-api-key-here"
      }
    }
  }
}

{
  "mcpServers": {
    "hopx-sandbox": {
      "command": "uvx",
      "args": ["hopx-mcp"],
      "env": {
        "HOPX_API_KEY": "your-api-key-here"
      }
    }
  }
}

Usage

Quick Execution

execute_code_isolated(
    code='print("Hello, World!")',
    language='python',
    timeout=30
)

Returns:

{
    'stdout': 'Hello, World!\n',
    'exit_code': 0,
    'execution_time': 0.123,
    'sandbox_id': '1762778786mxaco6r2'
}

Persistent Sandbox

For multiple operations in the same environment:

# Create sandbox
sandbox = create_sandbox(
    template_id='code-interpreter',
    timeout_seconds=3600
)

sandbox_id = sandbox['id']

# Run operations
execute_code(sandbox_id, 'import pandas as pd')
execute_code(sandbox_id, 'df = pd.read_csv("data.csv")')
file_write(sandbox_id, '/workspace/output.txt', 'results')
content = file_read(sandbox_id, '/workspace/output.txt')

# Cleanup
delete_sandbox(sandbox_id)

Available Tools

Supported Languages

• Python 3.11+: pandas, numpy, matplotlib, scipy, scikit-learn, requests
• JavaScript/Node.js 20: Standard libraries
• Bash: Unix utilities, git, curl, wget
• Go: Compilation support

Environment Variables

Required:

HOPX_API_KEY=your-api-key

Optional:

HOPX_BASE_URL=https://api.hopx.dev  # default

Troubleshooting

"401 Unauthorized"

API key not set or invalid.

Solution:

echo $HOPX_API_KEY  # Verify key is set

"Template not found"

Invalid template name.

Solution:

templates = list_templates(limit=20)  # Browse available templates

Slow First Execution

Container initializes in ~3 seconds after creation.

Cause: VM authentication setup Solution: Subsequent operations are immediate

Limitations

• Synchronous execution: Max 300 seconds
• Sandbox lifetime: Default 10 minutes (configurable)
• Template-specific language support

Security

Protected:

• Local system isolated from container execution
• Containers isolated from each other
• Auto-cleanup prevents resource leaks
• JWT authentication per sandbox

Considerations:

• Containers have internet access by default
• Code executed in HOPX cloud
• Follow your security policies for sensitive data

🔧 Technical Details

Built with:

• hopx-ai SDK v0.2.7
• FastMCP framework
• Model Context Protocol

📄 License

This project is under the MIT License. See LICENSE for details.

HOPX Terms of Service: hopx.ai/terms

Website | Documentation | GitHub