These are My 3 Must-Have MCP Servers for AI Coding

Key Concepts:

• MCP Servers (Machine Communication Protocol Servers): Servers facilitating communication between AI coding assistants and external resources.
• RAG (Retrieval-Augmented Generation): An AI framework that combines information retrieval with text generation.
• Pydantic AI: An AI agent framework.
• Crawl for AI: A free, open-source MCP server for providing up-to-date documentation.
• Brave: A web search engine used as an MCP server for accessing additional resources.
• Neon: A serverless Postgres platform used as an MCP server for database management with natural language.
• Windsurf/Cursor: AI coding assistants.

Three Essential MCP Servers for AI Coding Assistants

The video emphasizes the importance of three types of MCP servers when using AI coding assistants like Windsurf or Cursor. These servers enhance the capabilities of the AI by providing access to documentation, web search, and database management.

1. Documentation MCP Server (Crawl for AI)

• Purpose: Provides up-to-date documentation to the AI coding assistant.
• Example: The presenter uses "Crawl for AI," a free and open-source MCP server they built.
• Functionality: Allows the AI to access relevant documentation for code generation and problem-solving.
• Benefit: Ensures the AI uses the latest information, improving accuracy and reducing errors.

2. Web Search MCP Server (Brave)

• Purpose: Enables the AI coding assistant to search the web for additional resources.
• Example: The presenter uses the Brave search engine as an MCP server.
• Functionality: Allows the AI to find relevant code snippets, libraries, or solutions to specific problems.
• Benefit: Expands the AI's knowledge base beyond its internal training data.

3. Database Management MCP Server (Neon)

• Purpose: Allows the AI coding assistant to manage databases using natural language.
• Example: The presenter uses Neon, a serverless Postgres platform.
• Functionality: Enables the AI to create projects, define SQL tables, and manage records through natural language commands.
• Benefit: Simplifies database interactions, making it easier for the AI to build and manage applications.

Demonstration: Building an AI Agent with Windsurf and MCP Servers

The video demonstrates how these MCP servers work together in a practical scenario: building an AI agent using Windsurf and Pydantic AI.

• Process:
  1. The presenter asks Windsurf to build an AI agent with Pydantic AI.
  2. Windsurf leverages the Crawl for AI MCP server to access up-to-date documentation for Pydantic AI.
  3. Windsurf uses the Brave MCP server to search for additional resources if needed.
  4. Windsurf utilizes the Neon MCP server to create a new project and define the necessary SQL tables for the AI agent.
• Outcome: Windsurf generates the complete code for the AI agent, including an interface.
• Result: The AI agent is able to answer questions by leveraging its knowledge base, demonstrating the effectiveness of the MCP servers.

Conclusion

The video argues that using these three types of MCP servers significantly enhances the capabilities of AI coding assistants. By providing access to documentation, web search, and database management, these servers enable AI to generate more accurate, efficient, and complex code. The demonstration with Windsurf and Pydantic AI illustrates the practical benefits of integrating these MCP servers into the AI coding workflow.