XiYan MCP Server is a powerful MCP server designed to connect AI models with databases using natural language. Leveraging XiYan-SQL, a state-of-the-art text-to-SQL model, it allows developers to query databases with simple natural language, eliminating the need for complex SQL coding. It supports various LLMs, including GPT and Qwen, and offers both remote and local deployment options for enhanced security.

Key features include the ability to list available tables as resources and retrieve table content, providing AI models with the necessary context for accurate data retrieval. By translating natural language into SQL queries, XiYan MCP Server streamlines data access for AI, making it ideal for building intelligent data assistants and applications. Installation is straightforward via pip, with flexible configuration options for different LLMs and database systems like MySQL and PostgreSQL.

Natural Language Data Retrieval

XiYan MCP Server empowers AI models to retrieve data from databases using natural language queries, leveraging the power of XiYan-SQL, a state-of-the-art text-to-SQL model. This eliminates the need for users to write complex SQL queries, making data access more intuitive and accessible. The server translates natural language questions into SQL queries, executes them against the configured database, and returns the results to the AI model. This allows AI models to answer questions, generate reports, or perform other data-driven tasks based on real-time information from the database.

For example, a user could ask "What are the top 5 products by sales in the last month?" The XiYan MCP Server would translate this into an appropriate SQL query, execute it against the sales database, and return the list of products and their sales figures to the AI model. This feature significantly enhances the ability of AI models to interact with and understand structured data.

Supports Multiple LLMs and Databases

The server supports a variety of Large Language Models (LLMs), including general-purpose models like GPT and Qwen, as well as specialized text-to-SQL models like XiYanSQL-QwenCoder. This flexibility allows users to choose the model that best suits their needs and resources. Additionally, the server supports both MySQL and PostgreSQL databases, providing compatibility with a wide range of existing database infrastructures. This broad compatibility ensures that the XiYan MCP Server can be easily integrated into diverse environments.

For instance, a company using a PostgreSQL database and preferring the Qwen LLM can configure the server to use these technologies seamlessly. The server's adaptability to different LLMs and databases makes it a versatile tool for various data-driven applications. The configuration is managed through a YAML file, allowing for easy switching between different models and database connections.

Secure Local Mode Operation

XiYan MCP Server offers a pure local mode, enhancing data security and privacy. In this mode, the text-to-SQL translation and database querying are performed entirely on the user's machine, eliminating the need to send sensitive data to external APIs or cloud services. This is particularly important for organizations that handle confidential information and require strict data governance. The local mode ensures that data remains within the organization's control, reducing the risk of data breaches or unauthorized access.

To use the local mode, users need to install additional Python packages and download the necessary model files. The server then runs locally, processing natural language queries and retrieving data from the database without any external communication. This feature provides a high level of security and control over data, making it suitable for sensitive applications.