Memobase

🚀 MemoDB MCP Server

The MCP Server is designed to manage the conversation context and personal knowledge base of AI applications. It provides tools for user data, conversation content, and knowledge management through the Model Context Protocol (MCP).

✨ Features

1. User Management

• create_user: Create a new user.
• get_user: Retrieve user information.
• update_user: Update user information.
• delete_user: Delete a user.

2. Conversation Data Management

• insert_blob: Insert conversation data.
• get_blob: Retrieve conversation data.
• delete_blob: Delete conversation data.

3. Knowledge Base Management

• query_knowledge: Query the knowledge base.
  • Supports full-text search.
  • Supports filtering by type, tag, and source.
  • Supports limiting the number of returned results.
• add_knowledge: Add new knowledge.
  • Supports setting the knowledge source.
  • Supports setting the knowledge type.
  • Supports adding tags.
• update_knowledge: Update existing knowledge.
  • Supports updating content and metadata.
  • Supports modifying tags.
• relate_knowledge: Create knowledge associations.
  • Supports setting the association type.
  • Supports setting the association weight.

🔧 Technical Details

1. Type Safety

   • Implemented in TypeScript.
   • Complete type definitions and checks.
   • Compile-time error detection.

2. Error Handling

   • Comprehensive error handling mechanism.
   • Detailed error information.
   • Error logging.

3. API Design

   • Based on the JSON-RPC 2.0 protocol.
   • RESTful API style.
   • Clear interface definitions.

4. Scalability

   • Modular design.
   • Plugin-style tool registration.
   • Easy to add new features.

📦 Installation

1. Install dependencies:

npm install

2. Configure environment variables: Create a .env file and set the following:

MEMOBASE_API_URL=your_api_url    # API server address
MEMOBASE_API_KEY=your_api_key    # API access key

3. Build the project:

npm run build

4. Run the server:

# Production environment
npm start

# Development environment
npm run dev

💻 Usage Examples

Basic Usage

1. Add Knowledge

const result = await callTool('add_knowledge', {
  uid: 'user123',
  content: 'Artificial intelligence is a branch of computer science...',
  metadata: {
    source: 'wiki',
    type: 'article',
    tags: ['AI', 'Computer Science', 'Technology']
  }
});

2. Query Knowledge

const result = await callTool('query_knowledge', {
  uid: 'user123',
  query: 'Artificial intelligence',
  filters: {
    types: ['article'],
    tags: ['AI'],
    sources: ['wiki']
  },
  limit: 10
});

3. Relate Knowledge

const result = await callTool('relate_knowledge', {
  uid: 'user123',
  source_kid: 'knowledge1',
  target_kid: 'knowledge2',
  relation_type: 'related_to',
  weight: 0.8
});

📚 Documentation

1. Adding New Tools

• Add new functional modules to the code.
• Follow the existing design patterns and naming conventions.
• Ensure interface compatibility between modules.

2. Modifying Configuration

• Update the parameter values in the .env file.
• Keep the environment variable names consistent.
• Test the running effects under different configurations.

3. Running Tests

• Use npm test to execute unit tests.
• Check the test coverage report.
• Regularly update test cases.

🤔 Frequently Asked Questions

• How to get help? Refer to the documentation or contact the support team.
• How to handle errors? Check the log files and refer to the troubleshooting guide.

📄 License

This project is licensed under the [License Name] license. For more details, please refer to the LICENSE file.