Major Support for MCP

Hey, welcome to this video. Over the past year the Model Context Protocol became a common standard in the AI space and the big reason was Entropics' choice in late 2024 to open source MCP. It all started with the introduction post of the Model Context Protocol which you can read on the official Entropic website. The documentation for Model Context Protocol you can find on modelcontextprotocol.io. It's rich with diagrams, explanation, core concepts, SDKs and specifications. Popular languages are supported like Python, TypeScript and so on. So you find for example a TypeScript SDK as open source project on GitHub which means you can easily integrate or create your own MCP server even your clients and you find a lot of servers already on GitHub and on different registers on services. And this leads many developers to adopt MCP fast. Entropics' Cloudy Desktop used MCP early and tools like Kerza, WinServe and even VS Code has integrated and supported MCP latest in early 2025. And OpenAI create their own SDK for it with function calling but in March 2025 they officially adopted and support now MCP in their Agent SDK which you can find on openai.github.io. They have already their own page here which you can see here, Model Context Protocol and they support it also in their chat GBD desktop app. Microsoft introduced Model Context Protocol in March and supported it and integrated it also in their VS Code. So in their Code Editor and the last contender is Google. They announced recently the Agent2Agent protocol but Agent2Agent is more like a connection between MCP clients and hosts so it complements MCP which you can read here. Agent2Agent is an open protocol that complements Entropics' Model Context Protocol which provides helpful tools. How can this work? In real world setup you have here MCP hosts and the clients which talks to the servers, to the multiple servers. So you have multiple agents like a finance agent, a sales agent, marketing agent, whatever you can think of and these are connected with the Agent2Agent protocol from Google. So this is a setup you can think of. Pretty cool, this makes it more complete. And for me personally this means we can more safely focus now on building services and servers around this standard. And yeah this move solidifies MCP as the standard for connecting AI models to external tools and data sources. Now we are breaking down what MCP is and how it helps connect AI to different services and by the end you will see how MCP makes integration simpler, lowers the setup time and helps teams to build more connected AI tools.