Model Context Protocol (MCP)

About

Introduction

MCP, the Model Context Protocol, is an open protocol that enables seamless integration between LLM applications and external data sources and tools.

MCP is sometimes described as “OpenAPI for LLMs” or as “USB-C port for AI”, providing a uniform way to connect LLMs to resources they can use.

Details

The main entities of MCP are prompts, resources, and tools. MCP clients call MCP servers, either by invoking them as a subprocess and communicating via Standard Input/Output (stdio), Server-Sent Events (sse), or HTTP Streams (streamable-http), see transports.

Discuss

To get in touch with us to discuss CrateDB and MCP, head over to GitHub at Model Context Protocol (MCP) @ CrateDB or the Community Forum.

Usage

You can use MCP with CrateDB and CrateDB Cloud, either by selecting the CrateDB MCP Server suitable for Text-to-SQL and documentation retrieval, or by using community MCP servers that are compatible with PostgreSQL databases.

CrateDB MCP Server

The CrateDB MCP Server, available on PyPI and popular community hubs.

CrateDB MCP Server
Community MCP Servers

MCP servers mostly compatible with both PostgreSQL and CrateDB.

Community MCP Servers

To use an MCP server, you need a client that supports the protocol. The most notable ones are ChatGPT, Claude, Cline, Cursor, GitHub Copilot, Mistral AI, OpenAI Agents SDK, VS Code, Windsurf, and others.