Thanks for the great article! What I'm wondering is how a client, e.g. Cursor, relates a statement to one of the connected MCP servers? For example, in the case you've provided, if I type in "Read from the browser console logs..." How does it know it should use the "AgentDesk BrowserTools MCP"? Is it based on certain keywords? If yes, what happens if words like "console" or "browser" are also keywords in another MCP?
The best read's I had so far on MCP server's 🙌
Wow, thank you so much, Abdul! 🙏
Thank you for sharing. I didn’t know there was a term called MCP. I’m going to practice it soon.
Glad you liked, Thinh! Thank you!
Great article Jordan, super interesting and well written as always :)
And thanks for sharing Shirvu’s article, really enjoyed it too!
Thanks so much, Anton 🙏 glad you like!
Clear explanation. MCP looks simple at the protocol level, but production implementation is where things get interesting.
I wrote a practical series covering that next layer: https://substack.com/@andriitkachuk/note/c-284726045?utm_source=notes-share-action&r=5ozxhv
Thanks for the great article! What I'm wondering is how a client, e.g. Cursor, relates a statement to one of the connected MCP servers? For example, in the case you've provided, if I type in "Read from the browser console logs..." How does it know it should use the "AgentDesk BrowserTools MCP"? Is it based on certain keywords? If yes, what happens if words like "console" or "browser" are also keywords in another MCP?
Good one. Just sharing my own piece from Feb.
https://www.stackgazer.com/p/model-context-protocol-ai-development
It's great. Thanks for your session, which helped me understand and feel confident about using MCP for my project.