Model Context Protocol is how AI Agents get things done using Large Language Models.
Model Context Protocol is how AI Agents get things done using Large Language Models.
Hi I created PI MCP Server internally as a personal experiment and results are very good. we have done this as testing framework with agents and its not productized. If you need any help on this let me know.
Hi I created PI MCP Server internally as a personal experiment and results are very good. we have done this as testing framework with agents and its not productized. If you need any help on this let me know.
That is very interesting. I've been watching YouTube videos and reviewing the MCP Specification and, as a PI System Admin, I thought this would be a great way to expose all the data we've collected and stored in the PI System to an AI Agent that can be used to analyze the data by it asking questions.
Does your MCP Server use the PI Web API?
The alternative to writing an MCP Server from scratch would be to use a tool like Tatsoft Frameworx which, I think, gets me 70-80% of the way there.
Are you familiar with Tatsoft Frameworx?
Yes I used PI web API. I just used python libraries to create functions and these will be exposed as tools with Claude desktop or any other client.