How it works
A service runs an MCP server that describes the tools it offers, such as searching products or checking a URL, with their inputs and outputs. An AI application that supports MCP connects to the server and lets the model call those tools during a conversation, returning real data instead of guessing.
Why sites publish one
As people work through assistants, a product that an assistant can use directly becomes part of those workflows. Some sites also publish a small mcp.json file describing where their MCP server is, so tools can discover it, much as robots.txt and llms.txt describe a site to crawlers.
Before you build one
Start with the tasks users would actually ask an assistant to do with your product, expose only those, and protect anything that changes data behind authentication.
MCP compared with an API
An API is built for developers who write code against it. An MCP server wraps actions in a form an AI model can discover and call on its own, with descriptions the model reads to decide when to use each tool. Many services build an MCP server on top of their existing API.
Example
A launch platform's MCP server offers a search tool, so an assistant asked for new AI writing tools can query the platform's live listings and cite them.
Questions people ask
- Do I need an MCP server for my startup?
- Only if users would benefit from an assistant using your product directly. For most sites, a clear llms.txt and readable pages come first.
- Is MCP only for one AI company's products?
- No. It is an open protocol, and it is supported by several AI applications and developer tools.
Related terms
- llms.txtllms.txt is a Markdown file at the root of a website that gives large language models a short summary of the site and links to its most useful pages, so they can read it without parsing every page.
- ai.txtai.txt is a plain text file at a site's root that states, in plain language, how AI systems may use the site's content and which facts they should know about it.
- Generative Engine Optimization (GEO)Generative Engine Optimization (GEO) is the practice of making your content easy for AI assistants such as ChatGPT, Claude, Perplexity and Google's AI Overviews to find, understand and cite in their answers.
- AI citationAn AI citation is a link or mention of a source in an answer from an AI assistant such as ChatGPT, Perplexity, Claude or Google's AI Overviews, shown so readers can check where the answer came from.