302AI Sandbox MCP Server

🚀 302AI Sandbox MCP Server

A Model Context Protocol (MCP) service that supports the secure execution of arbitrary code by AI assistants, equipped with a code sandbox feature.

Chinese | English | Japanese

🚀 Quick Start

Preview

Here are some usage examples:

✨ Features

• Dynamic Tool List Loading: Supports real-time updates and loading of new tools.
• Multiple Operating Modes:
  • Local Standard Input Mode: Interact with the service via standard input.
  • Remote HTTP Server Mode: Provides RESTful API interfaces for convenient remote calls.

Tool List

Here are some main functions and service interfaces:

1. Code Execution
   • /execute: Execute a given code snippet.
2. Sandbox Management
   • /sandbox/create: Create a new sandbox environment.
   • /sandbox/list: List all existing sandbox instances.
   • /sandbox/destroy: Destroy a specified sandbox instance.
3. File Operations
   • /file/upload: Upload a file to the sandbox environment.
   • /file/download: Download a file from the sandbox environment.
4. Log Management
   • /log/view: View the sandbox's log output.
5. Resource Monitoring
   • /monitor/cpu: Get the CPU usage.
   • /monitor/memory: Get the memory usage.

📦 Installation

1. Install Dependencies:

   npm install
   

2. Build the Service:

   npm run build
   

3. Run the Service:

   npm start
   

For convenient debugging, the project provides the MCP Inspector tool:

npm run inspector

📚 Documentation

Configuration

Configure via Claude Desktop

Add the following content to config.json:

{
  "mcp_server": {
    "type": "claudedesktop",
    "endpoint": "http://localhost:1234"
  }
}

Configure via Cherry Studio

Add the following content to config.json:

{
  "mcp_server": {
    "type": "cherrystudio",
    "api_key": "your_api_key_here"
  }
}

Configure via ChatWise

Add the following content to config.json:

{
  "mcp_server": {
    "type": "chatwise",
    "endpoint": "http://localhost:5678"
  }
}

Debugging

Since the MCP server communicates via standard input and output, debugging can be challenging. We recommend using MCP Inspector:

npm run inspector

The Inspector will provide a browser address to access the debugging tool.

About 302.AI

302.AI is an enterprise-oriented AI application platform that offers pay-as-you-go services, ready-to-use solutions, and an open-source ecosystem. ✨

1. 🧠 Integrates the latest and most comprehensive AI capabilities, including but not limited to language models, image models, speech models, and video models.
2. 🚀 Develops in-depth applications based on foundational models - we develop real AI products, not just simple chatbots.
3. 💰 No monthly fees. All features are pay-as-you-go, fully open, truly achieving low barriers and high potential.
4. 🛠 A powerful team management backend, suitable for teams and small and medium-sized enterprises. One person can manage multiple users.
5. 🔗 All AI capabilities provide API access rights. All tools are open-source and customizable (in progress).
6. 💡 A strong development team launches 2 - 3 new applications every week, and the products are continuously updated. Developers interested are welcome to contact us.

⚠️ Important Note

Please ensure that you comply with relevant laws, regulations, and platform rules when using this service. Especially when dealing with sensitive information and performing high-risk operations, carefully evaluate their security and compliance.