Remix clone Hacker News

new | show | ask | jobs Github

▲

runako 13 days ago

I'm still trying to grok MCP, would be awesome if you could include usage examples in the doc.

Good luck!

▲

mparis 13 days ago | parent | next [-]

My first foray into using MCP was via Claude Desktop. Would be great if you packaged your tool such that one could add it with a few lines in their ‘~/Library/Application Support/Claude/claude_desktop_config.json’

	▲	jamestimmins 13 days ago \| parent [-]
		Same here. Tonight I added Whatsapp to Claude Desktop via Luke Harries' https://github.com/lharries/whatsapp-mcp. Very solid intro into how it all works.

▲

saberience 12 days ago | parent | prev | next [-]

It's not complicated at all.

All it does is expose methods as a "tool" which is then brought back to your LLM and defined with its name, description and input parameters.

E.g. Name: "MySqlTool", Description: "Allows arbitrary MySQL queries to the XYZ database", Parameters: "string: sqlToExecute"

The MCP Client (e.g. Claude Desktop, Claude Code), is configured to talk to an MCP server via stdio or sse, and calls a method like "tools/list", the server just sends a list back (in JSON) of all the tools, names, descriptions, params.

Then, if the LLM gets a query that mentions e.g. do a web search, or a web scraping, etc, it just outputs a tool use token then stops inferencing. Then the code calls that tool via stdio/sse (json-rpc), to the MCP server, which just runs that method, returns the result, then its added to the message history in the LLM, then inferencing runs again from the beginning.

▲

runako 12 days ago | parent [-]

I think people who have been building with LLMs have a different view on what is complicated vs not :-)

It may be easy for you to configure, but you dropped some acronyms in there that I would have to look up. I have definitely not personally set up anything like this.

▲

bavell 12 days ago | parent [-]

It's basically a simple rpc server, there's nothing complicated going on...

	▲	runako 12 days ago \| parent [-]
		Then please write up and share here some example usage documentation for someone who has never used MCP. (That was my suggestion upthread.) As a side note, do people here not realize that less complicated examples are often better for learning? Have we as a community forgotten this basic truism? Since it’s not complicated, you should be able to write it up quickly, and parallel commmenters to mine suggest there is an audience for such documentation. Thanks!

▲

romanovcode 12 days ago | parent | prev | next [-]

It's very simple and this is actually good example.

1. You add this MCP to your DB (make sure it is securely connected to your AI of choice of course)

2. Ask anything about your data, ask to make graphs, ask to make scheduled tasks, ask to analyze queries and show optimizations and so on.

3. Profit, literally. No need to pay BI companies thousands each month.

▲

teaearlgraycold 12 days ago | parent | prev [-]

The main things that made MCP hard for me to understand at first is that it’s both transport agnostic (so no leveraging semantic HTTP) and is an async task management protocol as well as a tool use protocol. The name itself is also poorly chosen. I would call it Tool Use Protocol.

Think about each MCP implementer like an agent’s input/output device.