Introduction to MCP

What is the Model Context Protocol?

The Model Context Protocol (MCP) is an open-source framework designed to bridge the gap between structured data and large language models (LLMs). Instead of relying on prompts or traditional SQL interfaces, MCP introduces a smarter way for LLMs to discover, understand, and work with your data by using registered tools and metadata.

Developed and maintained by Anthropic, MCP gives AI tools a structured understanding of the capabilities available for a given data source.

Introducing CData Code Assist MCP

The CData Code Assist MCP for DB2 is a lightweight local server that makes DB2 data accessible to MCP clients (such as Claude Desktop) using the MCP standard. Once configured, the MCP client can communicate with the MCP server to get the data it needs to answer nuanced questions like “What are the top 5 deals by revenue this quarter?” and it provides live answers without writing any SQL or teaching it what to do.

The CData Code Assist MCP for DB2 bundles your data connector with the CData Code Assist MCP and exposes a set of callable tools to MCP clients. These tools are:

Registered using the included CData Code Assist Configuration Tool.
Launched by the MCP client automatically.
Defined and described in Java files that MCP clients interpret directly.

Why This Approach?

Traditional prompt-based LLM workflows often run into token limits before the question is even asked. MCP solves this by using structured tool metadata and schemas.

No prompt engineering required.
No token waste.
No brittle, hand-coded logic.
You only need to define the connection, and the MCP Client handles the rest.

How It Works

After installing the CData Code Assist MCP and saving your connection:

CData Code Assist MCP registers your connector, loads its capabilities, and exposes the tools to the MCP Client.
The MCP client discovers the connection automatically and interprets the available tools through metadata.
When you ask a question, the MCP client determines the best tool to use and issues the request.

Once connected, the MCP client can answer your questions using live data from DB2. It securely sends requests to the CData Code Assist MCP, processes the results, and responds in natural language without any scripting or query writing required.

Designed for LLMs

MCP is a new way to connect AI to data. It is designed specifically for LLMs, using structured tool definitions instead of traditional query interfaces. Each tool is described with LLMs in mind so MCP clients don't rely on guessing or parsing output. It interprets structured definitions and uses them to generate intelligent, contextual queries behind the scenes.

Why CData?

The CData Code Assist MCP for DB2 shows what's possible when you combine industry-leading data connectivity with a modern AI interface. As LLMs evolve, CData is helping users move beyond traditional query interfaces into intuitive, intelligent access.

Code Assist MCP for DB2