Mindbridge MCP

🚀 MindBridge MCP Server Introduction

MindBridge MCP Server is a powerful multi-vendor AI interface management platform. It supports multiple mainstream LLM providers such as OpenAI, Anthropic, Google, DeepSeek, and Ollama, as well as custom OpenAI-compatible services. The platform offers inference optimization features, enabling users to manage and invoke different AI models more efficiently.

🚀 Quick Start

The following steps will guide you through setting up and running the MindBridge MCP Server.

📦 Installation

Installation Steps

1. Clone the repository

   git clone https://github.com/pinkpixel/mindbridge.git
   cd mindbridge
   

2. Install dependencies

   npm install
   

3. Start the development environment (optional)

   npm run dev
   

4. Build for production

   npm run build
   npm start
   

Configuration Options

MindBridge can be configured in the following ways:

1. Command-line arguments

2. Environment variables

   • MB_API_KEY: OpenAI API key
   • MB_OPENROUTER_KEY: OpenRouter access key
   • MB_DEEPSEEK_KEY: DeepSeek API key
   • MB_GOOGLE_PROJECT_ID: Google Cloud project ID
   • MB_Ollama_HOST: Ollama server address

3. Configuration file

   {
     providers: {
       openai: {
         apiKey: 'your-openai-key',
         models: ['gpt-4', 'gpt-3.5']
       },
       google: {
         projectId: 'your-google-project-id',
         models: ['gemini-pro']
       }
     }
   }
   

✨ Features

1. Multi-vendor Support

• OpenAI: Supports models like GPT-4 and GPT-3.5.
• Anthropic: Supports Claude 2 and 3 series models.
• Google: Integrates Gemini models.
• DeepSeek: Supports DeepSeek series models.
• Ollama: Supports locally run open-source models.

2. Inference Optimization

• Inference intensity: Set the inference effort level (low/medium/high).
• Context length: Adjust the maximum context window.
• Temperature parameter: Control the output randomness.

3. Out-of-the-box Features

• Get multiple opinions: Use the getSecondOpinion method to get responses from multiple models.
• Vendor list: List all configured LLM providers and their available models.
• Inference model list: View optimized models suitable for inference tasks.

💻 Usage Examples

Basic Usage

mindBridge.getCompletion({
  provider: 'openai',
  model: 'gpt-4',
  prompt: '简单介绍一下人工智能的基本概念？'
}).then(console.log);

Advanced Usage

Custom Parameters

mindBridge.getCompletion({
  provider: 'anthropic',
  model: 'claude-3',
  prompt: '分析一下深度学习的发展历程',
  temperature: 0.7,
  maxTokens: 500
}).then(console.log);

Use Custom Model

mindBridge.getCompletion({
  provider: 'custom-openai',
  model: 'my-custom-model',
  prompt: '欢迎使用我的自定义模型！'
}).then(console.log);

📚 Documentation

Contribute Code

1. Local development

   npm run dev
   

2. Unit tests

   npm test
   

3. Submit code

   • Format code before submission: npm run lint
   • Check code quality: npm run lint:fix

Other Contributions

• Fix issues.
• Add new features.
• Improve documentation.
• Provide feedback.

Project Maintainers

• Zhang San (GitHub: @zhangsan)
• Li Si (GitHub: @lisi)

📄 License

MIT License

Contact Information

If you have any questions or suggestions, please send an email to support@mindbridge.com or submit an issue on GitHub.