diff --git a/ai/model-context-protocol.mdx b/ai/model-context-protocol.mdx index a40f30bb4..a0a7df054 100644 --- a/ai/model-context-protocol.mdx +++ b/ai/model-context-protocol.mdx @@ -10,11 +10,11 @@ import { PreviewButton } from "/snippets/previewbutton.jsx" The Model Context Protocol (MCP) is an open protocol that creates standardized connections between AI applications and external services, like documentation. Mintlify generates an MCP server from your documentation, preparing your content for the broader AI ecosystem where any MCP client like Claude, Cursor, Goose, ChatGPT, and others can connect to your documentation. -Your MCP server exposes tools for AI applications to search your documentation and retrieve full page content. Your users must connect your MCP server to their tools. +Your MCP server exposes tools for AI applications to search your documentation and read page content directly from your documentation source files. Your users must connect your MCP server to their tools. ### How MCP servers work -When an AI application connects to your documentation MCP server, it can search your documentation and retrieve full page content directly in response to a user's prompt instead of relying on information from its training data or making a generic web search. Your MCP server provides access to all indexed content on your documentation site. +When an AI application connects to your documentation MCP server, it can search your documentation and read page content directly in response to a user's prompt instead of relying on information from its training data or making a generic web search. Your MCP server provides access to all indexed content on your documentation site. - AI applications can proactively search your documentation while generating a response even if not explicitly asked to search your documentation for an answer. - AI applications determine when to use the available tools based on the context of the conversation and the relevance of your documentation. @@ -29,9 +29,9 @@ When an AI application connects to your documentation MCP server, it can search Your MCP server provides two tools that AI applications can use: - **Search**: Searches across your documentation to find relevant content, returning snippets with titles and links. Use this when you need to discover information or find pages matching a query. -- **Get page**: Retrieves the full content of a specific documentation page by its path. Use this when you already know the page path, such as from search results, and need the complete content rather than a snippet. +- **Query docs filesystem**: Runs a shell command against a virtualized, read-only filesystem that contains your documentation source files. Use this to read full page content, list the file tree, or search across multiple pages in a single call. Supported commands include `head`, `cat`, `tree`, `rg` (ripgrep), and other read-only utilities. -AI applications determine when to use each tool based on the context of the conversation. For example, an AI application might first search your documentation to find relevant pages, then use the get page tool to retrieve the full content of the most relevant result. +AI applications determine when to use each tool based on the context of the conversation. For example, an AI application might first search your documentation to find relevant pages, then use the filesystem tool to read the full content of those pages. ### Search parameters @@ -44,11 +44,11 @@ The MCP search tool supports optional parameters that AI applications use to con AI applications determine when to apply these parameters based on the context of the user's query. For example, if a user asks about a specific API version, the AI application may automatically apply the appropriate filter to provide more relevant results. -### Get page parameters +### Query docs filesystem parameters -The get page tool accepts one required parameter: +The query docs filesystem tool accepts one required parameter: -- **`page`**: The page path to retrieve, such as `'api-reference/create-customer'`. Use page paths returned from search results. +- **`command`**: A shell command to run against the virtualized documentation filesystem. For example, `tree / -L 2` to list the file structure, `head -80 /path/page.mdx` to read a page, or `rg -il "keyword" /` to search across all pages. ### MCP compared to web search diff --git a/es/ai/model-context-protocol.mdx b/es/ai/model-context-protocol.mdx index 8a691e97f..1f902be13 100644 --- a/es/ai/model-context-protocol.mdx +++ b/es/ai/model-context-protocol.mdx @@ -12,13 +12,13 @@ import { PreviewButton } from "/snippets/previewbutton.jsx" El Model Context Protocol (MCP) es un protocolo abierto que crea conexiones estandarizadas entre aplicaciones de IA y servicios externos, como la documentación. Mintlify genera un servidor MCP a partir de tu documentación, preparando tu contenido para el ecosistema de IA más amplio, donde cualquier cliente MCP (como Claude, Cursor, Goose y otros) puede conectarse a tu documentación. -Tu servidor MCP expone herramientas para que las aplicaciones de IA puedan buscar en tu documentación y obtener el contenido completo de las páginas. Tus usuarios deben conectar tu servidor MCP a sus herramientas. +Tu servidor MCP expone herramientas para que las aplicaciones de IA puedan buscar en tu documentación y leer el contenido de las páginas directamente desde tus archivos fuente. Tus usuarios deben conectar tu servidor MCP a sus herramientas.