MCP相关概念

马学

798人浏览 · 2026-02-27 10:59:04

马学 · 2026-02-27 10:59:04 发布

what is MCP?

Model context protocol (MCP) is a standardized interface and framework that

allows AI models to seamlessly interact with external tools, resources, and

environments.

MCP acts as a universal connector for AI systems to capabilities (tools, etc.), similar to how USB-C standardizes connections between electronic devices.

Why was MCP created?

The problem

M×N integration problem

The solution

work flow:

MCP Architecture Overview

three main roles to understand: the Host, the Client, and the Server

Host

The Host is the user-facing AI application, the environment where the AI model lives and interacts with the user.

Client

While the Host decides what to do, the Client knows how to speak MCP to actually carry out those instructions with the server.

Server

The MCP Server is the external program or service that actually provides the capabilities (tools, data, etc.) to the application.

An MCP Server can be thought of as a wrapper around some functionality, which exposes a set of actions or resources in a standardized way so that any MCP Client can invoke them.

The key is that the Server advertises what it can do in a standard format (so the client can query and understand available tools) and will execute requests coming from the client, then return results.

Tools, Resources and Prompts

Tools, prompts and resources form the three core capabilities of the MCP framework. Capabilities are essentially the features or functions that the server makes available.

Tools: Executable actions or functions that the AI (host/client) can invoke (often with side effects or external API calls).
Resources: Read-only data sources that the AI (host/client) can query for information (no side effects, just retrieval).
Prompts: Predefined prompt templates or workflows that the server can supply.

Tools

Tools are what they sound like: functions that do something on behalf of the AI model. These are typically operations that can have effects or require computation beyond the AI’s own capabilities.

Importantly, Tools are usually triggered by the AI model’s choice, which means the LLM (via the host) decides to call a tool when it determines it needs that functionality.

Resources

Resources provide read-only data to the AI model.These are like databases or knowledge bases that the AI can query to get information, but not modify.

Unlike tools, resources typically do not involve heavy computation or side effects, since they are often just information lookup.

Another key difference is that resources are usually accessed under the host application’s control (not spontaneously by the model). In practice, this might mean the Host knows when to fetch a certain context for the model.

For instance, if a user says, “Use the company handbook to answer my question,” the Host might call a resource that retrieves relevant handbook sections and feeds them to the model.

Resources could include a local file’s contents, a snippet from a knowledge base or documentation, a database query result (read-only), or any static data like configuration info.

Prompts

Prompts in the MCP context are a special concept: they are predefined prompt templates or conversation flows that can be injected to guide the AI’s behavior.

The user might pick a prompt/template from a UI (e.g., “Summarize this document” template), which the host then fetches from the server.

The model doesn’t spontaneously decide to use prompts the way it does tools.

Rather, the prompt sets the stage before the model starts generating. In that sense, prompts are often fetched at the beginning of an interaction or when the user chooses a specific “mode”.