Folks, I have been telling you for a year: keep an eye on MCP. It is the protocol proposed by Anthropic, which then passed under the Linux Foundation, because it lets you build services that AI agents can use. I have said it in my last few videos too: set the fear and the anxiety aside and start playing with the stuff that slowly slides into the background but stays constant. You can bet on it. MCP is one of those things.
The signal is getting loud
Model Context Protocol is sending a signal that comes through loud and clear: a lot of products in the Google suite now have an MCP version. There is a Google Maps MCP, a Google Drive MCP, a Google Ads MCP. So if you are not looking at MCP, I am going to smack every one of you on the back of the head. And of course, version 2 of the Cheshire Cat supports MCP natively.
The training I am putting together
One more thing: I am organizing some training on MCP. First an internal one for the Cheshire Cat core contributors, as a thank you for everything they have contributed over time, so the first one goes to them. After that I would like to do something a bit more public and educational about MCP, including how it hooks into the Cheshire Cat to build agents. Think agents that let you explore Google Maps and pop out widgets with the directions you need to take, or agents that talk to GitLab, to Google Drive, to the files on your system. And all of this obviously turns into a stream of content I put out online, plus a specific corporate course built around it.
So keep an eye on me, and above all tell me: is this stuff useful to you, or is it too advanced and I should just keep talking about the usual nonsense?
