{"id":19033,"date":"2025-07-07T15:06:51","date_gmt":"2025-07-07T18:06:51","guid":{"rendered":"https:\/\/loxias.ai\/?p=19033"},"modified":"2026-06-30T16:02:04","modified_gmt":"2026-06-30T15:02:04","slug":"mcp-the-protocol-thats-reshaping-ai-to-the-world","status":"publish","type":"post","link":"https:\/\/loxias.ai\/es\/blog\/mcp-the-protocol-thats-reshaping-ai-to-the-world\/","title":{"rendered":"MCP: the protocol that&#8217;s reshaping the way AI connects to the world"},"content":{"rendered":"\n<p class=\"wp-block-paragraph\"><span style=\"font-weight: 400;\">Artificial intelligence (AI) is moving beyond answering questions to performing real tasks\u2014and that\u2019s changing everything. With increasingly powerful models and the rise of AI agents, a new need has emerged: how can these intelligences be connected to external data sources, tools, and systems in a modular and scalable way?<\/span><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><span style=\"font-weight: 400;\">The answer comes in the form of a new technical standard: the <\/span><a href=\"https:\/\/www.anthropic.com\/news\/model-context-protocol\" rel=\"nofollow noopener\" target=\"_blank\"><b>Model Context Protocol<\/b><\/a><span style=\"font-weight: 400;\"> (MCP). Designed to be the \u201c<\/span><b>universal USB port<\/b><span style=\"font-weight: 400;\">\u201d (Figure 1) between large language models (LLMs) and the real world, the MCP is quickly becoming the invisible infrastructure behind intelligent agents, automated workflows, and enterprise AI applications.<\/span><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><span style=\"font-weight: 400;\">Figure 1 \u2013 MCP a \u201cuniversal USB port\u201d<\/span><\/p>\n\n\n\n<figure class=\"wp-block-image aligncenter size-full wp-image-19034\"><img fetchpriority=\"high\" decoding=\"async\" width=\"1024\" height=\"683\" src=\"http:\/\/loxias.ai\/wp-content\/uploads\/2025\/07\/Figura1-1024x683-1.webp\" alt=\"\" class=\"wp-image-20543\" srcset=\"https:\/\/loxias.ai\/wp-content\/uploads\/2025\/07\/Figura1-1024x683-1.webp 1024w, https:\/\/loxias.ai\/wp-content\/uploads\/2025\/07\/Figura1-1024x683-1-600x400.webp 600w, https:\/\/loxias.ai\/wp-content\/uploads\/2025\/07\/Figura1-1024x683-1-300x200.webp 300w, https:\/\/loxias.ai\/wp-content\/uploads\/2025\/07\/Figura1-1024x683-1-768x512.webp 768w\" sizes=\"(max-width: 1024px) 100vw, 1024px\"><figcaption class=\"wp-element-caption\">AI Generated Image<\/figcaption><\/figure>\n\n\n\n<h5 class=\"wp-block-heading\">\u00a0<\/h5>\n\n\n\n<h5 class=\"wp-block-heading\"><b>What is MCP?<\/b><\/h5>\n\n\n\n<p class=\"wp-block-paragraph\"><span style=\"font-weight: 400;\">MCP was launched by Anthropic in 2024 to address the complexity and scalability challenges of manually connecting AI applications to external APIs. It is an open protocol that standardizes the connection between large language models (LLMs) and external tools such as databases, APIs, file systems, ERPs, CRMs, and other types of systems. It eliminates custom integrations by creating a modular, interoperable channel for AI to access and manipulate real-world information.<\/span><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><span style=\"font-weight: 400;\">It is designed as a \u201cuniversal layer\u201d for connecting LLMs to any external resource in a secure, auditable, and extensible way.<\/span><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u00a0<\/p>\n\n\n\n<h6 class=\"wp-block-heading\"><b>How does it work? Understanding the architecture<\/b><\/h6>\n\n\n\n<p class=\"wp-block-paragraph\"><span style=\"font-weight: 400;\">The MCP is based on a <\/span><b>client-server architecture<\/b><span style=\"font-weight: 400;\"> and has three main blocks, as illustrated in Figure 2:<\/span><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><b>Host<\/b><span style=\"font-weight: 400;\">: the application where the model lives (such as a chatbot, dashboard, text editor). It incorporates one or more <\/span><b>MCP clients<\/b><span style=\"font-weight: 400;\">.<\/span><\/li>\n\n\n\n<li><b>MCP client<\/b><span style=\"font-weight: 400;\">: makes calls to <\/span><b>MCP servers<\/b><span style=\"font-weight: 400;\">. Can handle multiple servers simultaneously.<\/span><\/li>\n\n\n\n<li><b>MCP server<\/b><span style=\"font-weight: 400;\">: represents an external tool (like GitHub, Notion, browser, database, etc.) and exposes<\/span><b>: Recursos<\/b><span style=\"font-weight: 400;\">, <\/span><b>Herramientas<\/b><span style=\"font-weight: 400;\">, and <\/span><b>Prompts<\/b><span style=\"font-weight: 400;\">.<\/span><\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">\u00a0<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><span style=\"font-weight: 400;\">Figure 2 \u2013 MCP Architecture<\/span><\/p>\n\n\n\n<figure class=\"wp-block-image aligncenter size-full wp-image-19035\"><img decoding=\"async\" width=\"1024\" height=\"442\" src=\"http:\/\/loxias.ai\/wp-content\/uploads\/2025\/07\/Figura2-1024x442-1.webp\" alt=\"\" class=\"wp-image-20544\" srcset=\"https:\/\/loxias.ai\/wp-content\/uploads\/2025\/07\/Figura2-1024x442-1.webp 1024w, https:\/\/loxias.ai\/wp-content\/uploads\/2025\/07\/Figura2-1024x442-1-600x259.webp 600w, https:\/\/loxias.ai\/wp-content\/uploads\/2025\/07\/Figura2-1024x442-1-300x129.webp 300w, https:\/\/loxias.ai\/wp-content\/uploads\/2025\/07\/Figura2-1024x442-1-768x332.webp 768w\" sizes=\"(max-width: 1024px) 100vw, 1024px\"><figcaption class=\"wp-element-caption\">Source: Anthropic<\/figcaption><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">\u00a0<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u00a0<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><span style=\"font-weight: 400;\">Here is a more detailed description of the MCP Server components that are illustrated in Figure 3:<\/span><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><b>Recursos<\/b><span style=\"font-weight: 400;\"> (data): resources serve as a means of exposing data from the server to the client and can take many forms depending on the type of MCP server you are building. For example, in Playwright\u2019s MCP server that focuses on browser automation, resources might include console logs. In the case of a database MCP server such as SQLite, the resource might be the database file itself. And if you are building a Retrieval-Augmented Generation (RAG) server, the retrieved data might also be considered a resource.<\/span><\/li>\n\n\n\n<li><b>Herramientas<\/b><span style=\"font-weight: 400;\"> (executable functions): tools are the core components designed to be controlled by the model. They allow Large Language Models (LLMs) to perform actions on the MCP server. Tools are a powerful primitive within the Model Context Protocol (MCP) that allows servers to expose executable functionality to clients. For example, in a database MCP, tools include operations such as creating or selecting tables. In the context of GitHub MCP, tools might involve actions such as creating a commit or pulling a branch. In browser automation, tools might include navigating to a URL or clicking a button.<\/span><\/li>\n\n\n\n<li><b>Prompts<\/b><span style=\"font-weight: 400;\"> (instructions): prompts provide a method for exposing prompt templates from the server to the client, standardizing common interaction patterns. For example, on a GitHub server, prompts can be exposed to create a standard commit message from code. This simplifies the process because instead of writing prompts on the client side, server developers can create prompt templates for commonly needed interactions and expose them through a prompt artifact.<\/span><\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">\u00a0<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><span style=\"font-weight: 400;\">Figure 3 \u2013 MCP protocol operation<\/span><\/p>\n\n\n\n<figure class=\"wp-block-image aligncenter size-full wp-image-19036\"><img decoding=\"async\" width=\"1024\" height=\"697\" src=\"http:\/\/loxias.ai\/wp-content\/uploads\/2025\/07\/Figura3-1024x697-1.webp\" alt=\"\" class=\"wp-image-20545\" srcset=\"https:\/\/loxias.ai\/wp-content\/uploads\/2025\/07\/Figura3-1024x697-1.webp 1024w, https:\/\/loxias.ai\/wp-content\/uploads\/2025\/07\/Figura3-1024x697-1-600x408.webp 600w, https:\/\/loxias.ai\/wp-content\/uploads\/2025\/07\/Figura3-1024x697-1-300x204.webp 300w, https:\/\/loxias.ai\/wp-content\/uploads\/2025\/07\/Figura3-1024x697-1-768x523.webp 768w\" sizes=\"(max-width: 1024px) 100vw, 1024px\"><figcaption class=\"wp-element-caption\">Source: Anthrop\\c<\/figcaption><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">\u00a0<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><span style=\"font-weight: 400;\">Communication can occur via local <\/span><b>stdio<\/b><span style=\"font-weight: 400;\"> process (standard input output), or network streaming <\/span><b>HTTP + SSE<\/b><span style=\"font-weight: 400;\"> or <\/span><b>streamable HTTP<\/b><span style=\"font-weight: 400;\">.<\/span><\/p>\n\n\n\n<h6 class=\"wp-block-heading\">\u00a0<\/h6>\n\n\n\n<h6 class=\"wp-block-heading\"><b>Real and advanced use cases<\/b><\/h6>\n\n\n\n<ul class=\"wp-block-list\">\n<li><b>Blender + Claude Desktop: <\/b><span style=\"font-weight: 400;\">using Docker, it is possible to create MCP servers for tools like Blender. This allows, for example, an LLM to create or modify 3D scenes with natural language commands.<\/span><\/li>\n\n\n\n<li><b>Perplexity Windows: <\/b><span style=\"font-weight: 400;\">On Windows, MCP-enabled agents can access local files, terminal, web browsing, and applications, with controlled logs and permissions.<\/span><\/li>\n\n\n\n<li><b>Corporate Integrations: <\/b><span style=\"font-weight: 400;\">companies like Microsoft, AWS, IBM and startups like Triggo.ai already use MCP to integrate AI into proprietary databases with compliance.<\/span><\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">\u00a0<\/p>\n\n\n\n<h5 class=\"wp-block-heading\"><b>A powerful protocol, but not without risks<\/b><\/h5>\n\n\n\n<p class=\"wp-block-paragraph\"><span style=\"font-weight: 400;\">The more power we give to a model, the more security it needs. And with MCP, two risks come to the fore:<\/span><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><b>Prompt Injection: <\/b><span style=\"font-weight: 400;\">technique in which an attacker injects hidden commands into the text sent to the model. The agent can be manipulated to perform incorrect tasks, leak data, or disobey rules.<\/span><\/li>\n\n\n\n<li><b>Tool Poisoning: <\/b><span style=\"font-weight: 400;\">manipulation of tools connected to the model. If an MCP server is compromised, it can provide erroneous responses, execute dangerous commands, or act as a Trojan horse.<\/span><\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">\u00a0<\/p>\n\n\n\n<h5 class=\"wp-block-heading\"><b>Best practices for safe use of MCP<\/b><\/h5>\n\n\n\n<ul class=\"wp-block-list\">\n<li><b>Input validation<\/b><span style=\"font-weight: 400;\">: never blindly trust what comes from the model or the user.<\/span><\/li>\n\n\n\n<li><b>Audit and logs<\/b><span style=\"font-weight: 400;\">: log all calls and responses between MCP clients and servers.<\/span><\/li>\n\n\n\n<li><b>Permissions control<\/b><span style=\"font-weight: 400;\">: do not expose more tools than necessary.<\/span><\/li>\n\n\n\n<li><b>Cross-checking answers<\/b><span style=\"font-weight: 400;\">: use redundancy to avoid manipulated responses.<\/span><\/li>\n\n\n\n<li><b>Isolation<\/b><span style=\"font-weight: 400;\">: limit the scope of commands executed by tools.<\/span><\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">\u00a0<\/p>\n\n\n\n<h5 class=\"wp-block-heading\"><b>Conclusi\u00f3n<\/b><\/h5>\n\n\n\n<p class=\"wp-block-paragraph\"><span style=\"font-weight: 400;\">MCP is more than a protocol\u2014it\u2019s a strategic step toward making AI truly integrated, contextual, and autonomous. It transforms models into agents capable of acting in the real world with precision and responsibility. But with great power comes great responsibility: it\u2019s essential that MCP-based architectures are built with security, governance, and transparency in mind from the very first command.<\/span><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong><em>If you\u2019re building the future of AI, MCP will be your path. The question is: will you be ready?<\/em><\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u00a0<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><em>If you enjoy this content, maybe you would like to read more about\u00a0<a href=\"https:\/\/loxias.ai\/es\/blog\/rag-the-secret-behind-ais-to-business\/\" target=\"_blank\" rel=\"noopener\">RAG: The Secret Behind AIs That Truly Understand Your Business<\/a>.<\/em><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><em>Sign up for our newsletter and be the first to receive our latest articles. Stay informed on data analytics insights, business intelligence trends, marketing strategies, and the latest advancements in artificial intelligence.<\/em><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Artificial intelligence (AI) is moving beyond answering questions to performing real tasks\u2014and that\u2019s changing everything. With increasingly powerful models and the rise of AI agents, a new need has emerged: how can these intelligences be connected to external data sources, tools, and systems in a modular and scalable way? The answer comes in the form [&hellip;]<\/p>\n","protected":false},"author":7,"featured_media":20541,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"inline_featured_image":false,"footnotes":""},"categories":[48],"tags":[58,106,107,49,80,81,108],"class_list":["post-19033","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-artificial-intelligence","tag-ai","tag-anthropic","tag-apis","tag-artificial-intelligence","tag-large-language-models","tag-llms","tag-mcp"],"acf":[],"_links":{"self":[{"href":"https:\/\/loxias.ai\/wp-json\/wp\/v2\/posts\/19033","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/loxias.ai\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/loxias.ai\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/loxias.ai\/wp-json\/wp\/v2\/users\/7"}],"replies":[{"embeddable":true,"href":"https:\/\/loxias.ai\/wp-json\/wp\/v2\/comments?post=19033"}],"version-history":[{"count":1,"href":"https:\/\/loxias.ai\/wp-json\/wp\/v2\/posts\/19033\/revisions"}],"predecessor-version":[{"id":22380,"href":"https:\/\/loxias.ai\/wp-json\/wp\/v2\/posts\/19033\/revisions\/22380"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/loxias.ai\/wp-json\/wp\/v2\/media\/20541"}],"wp:attachment":[{"href":"https:\/\/loxias.ai\/wp-json\/wp\/v2\/media?parent=19033"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/loxias.ai\/wp-json\/wp\/v2\/categories?post=19033"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/loxias.ai\/wp-json\/wp\/v2\/tags?post=19033"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}