Frequently Asked Questions about MySQL MCP Server

Q: What is an MCP (Model Context Protocol) Server, and why should I use it instead of a regular API?

A: An MCP Server acts as a bridge between AI models and external tools or data sources, like a MySQL database. It uses a standardized protocol (MCP) developed by Anthropic. Compared to traditional APIs, MCP offers improved AI integration, automated permission management, and built-in security features, designed for AI assistants to understand and interact safely with external systems.

Q: What are the key benefits of using the MySQL MCP Server on the UBOS platform?

A: The MySQL MCP Server on UBOS provides secure and dynamic connections to MySQL databases, safe query execution (SELECT only), easy table structure viewing, built-in SQL injection protection, and dual deployment modes (stdio and HTTP/SSE). This allows AI assistants to safely query and analyze data without the risks of direct SQL manipulation.

Q: What database operations are supported by the MySQL MCP Server?

A: For security reasons, the MySQL MCP Server primarily supports SELECT queries. This read-only access ensures that AI cannot modify or delete data. It can show tables, describe table structures, and execute queries to retrieve data.

Q: What are the differences between stdio and HTTP/SSE modes? Which one should I choose?

A: stdio mode is simpler and more direct, requiring no port management, ideal for personal development and simple use cases. However, it needs to be restarted each time VSCode restarts. HTTP/SSE mode is persistent, supports auto-approval, but requires port management. It’s recommended for frequent use and team collaboration.

Q: How do I connect the MySQL MCP Server to my AI client or assistant?

A: You need to configure your AI client (e.g., Cline in VSCode) with the MCP Server details, including the connection command (for stdio) or the URL and type (for HTTP/SSE). Also, you need to set the host, port, username, password and database name of your MySQL database.

Q: Can I connect to a remote MySQL database using the MySQL MCP Server?

A: Yes, you can. When configuring the connection, simply specify the remote host address and the correct port number (usually 3306). Ensure that the remote MySQL server is configured to allow connections from your network.

Q: How secure is the MySQL MCP Server? What security measures are in place?

A: The MySQL MCP Server includes several security features: read-only access (SELECT queries only), SQL injection protection, operation limitations (no DROP, DELETE, UPDATE, INSERT), and parameter validation. It is also highly recommended to create a dedicated read-only database user for the MCP Server.

Q: How do I troubleshoot common issues like installation failures or connection problems?

A: Common troubleshooting steps include checking the execution policy (for PowerShell), verifying the dist/index.js file exists after building, ensuring the correct paths are used in configurations, checking database credentials, and ensuring the MySQL service is running. Detailed debugging steps are provided in the README.

Q: What if the default port (3100) is already in use? How can I change the port?

A: You can change the port by modifying the package.json script for start-gateway and specifying a different port. For example, change --port 3100 to --port 3200 and update the corresponding URL in your AI client configuration.

Q: How can I configure the MySQL MCP Server using environment variables?

A: Create a .env file in the project root to store sensitive information like database credentials. You can then access these variables in your configuration using process.env.VARIABLE_NAME. This prevents you from hardcoding sensitive data in your application.

Q: Does the MySQL MCP Server support other database types besides MySQL?

A: Currently, only MySQL is supported. However, the architecture is designed to be extensible, with future plans to support PostgreSQL, SQLite, SQL Server, and Oracle.

Q: What is UBOS, and how does it enhance the MySQL MCP Server?

A: UBOS is a full-stack AI Agent Development Platform designed to help you orchestrate AI Agents, connect them with your enterprise data, build custom AI Agents with your LLM model and Multi-Agent Systems. Using the MySQL MCP Server with UBOS gives you access to a broader ecosystem of AI development tools and simplifies managing AI-driven database interactions.