Introduction
MCP stands for Model Context Protocol.
In simple terms, Model Context Protocol (MCP) is an open standard protocol. Its main objectives are:
Standardized connectivity: To provide large language models (LLMs) with a standardized way to connect to and use external tools, data sources, or systems (such as databases, APIs, software, etc.). Enhanced model capabilities: Through this standardized interface, LLMs can go beyond the limitations of their internal knowledge base and training data, access real-time information, and perform specific operations (e.g., sending emails, querying orders), thereby significantly enhancing their functionality and application scenarios. Ecosystem development: As an open standard, MCP aims to make it easier for different developers and platforms to develop and integrate plugins or tools for LLMs, lower the development barrier, and promote the prosperity of the intelligent agent (Agent) ecosystem.
You can think of MCP as a "universal adapter" or a "connector standard" that enables large models to more conveniently and consistently "invoke" functionalities from the external world, thus becoming more powerful and practical. At the same time, you can choose to integrate existing MCP services or independently develop your own dedicated MCP service based on your needs. For example, platforms such as ModelScope MCP Marketplace and Higress MCP Marketplace have already aggregated a wealth of service resources for you to explore and select, helping you quickly find solutions that fit your business scenarios and improve project development efficiency and innovation capabilities.
MCP Marketplaces
ModelScope MCP Marketplace (https://www.modelscope.cn/mcp): The ModelScope community has listed over a thousand popular MCP services, including new MCP services from Alipay, MiniMax, etc., providing AI developers with rich MCP services and debugging tools.
Higress MCP Marketplace (https://mcp.higress.ai/): The Higress MCP Marketplace has currently listed multiple popular MCP services.
Use Cases
Databases: Automatically retrieve schemas from databases to generate DAO code or perform data queries.
Online documents: Generate code, optimize, or provide intelligent Q&A based on the content of online documents.
Design systems: Generate front-end code from design drafts in design systems to reproduce front-end interactive designs.
Operating Modes
STDIO mode: In this mode, communication is conducted via standard input/output streams, and the service runs locally.
SSE mode: Communication is performed using the Server-Sent Events (SSE) protocol, and the service can run either remotely or locally.