-
Notifications
You must be signed in to change notification settings - Fork 18
Description
Hi folks!
I like the MCP initiative! The following information on MCPStack might be of interest. In a nutshell, if you are familiar with Scikit-learn, MCPStack will let you build a successful chain of MCPs that resembles a Scikit-learn pipeline (instead of preprocessor, transformers, and estimators; you have MCPs); You can then execute this into any output, like Claude Desktop, or any other possibility (STDIO via FastMCP, etc.).
For the time being, we have two MCPs available (see marketplace) mainly because they were investigated prior to being mirrored in MCP-Pipeline (the org) by a team during one of their projects & papers at MIT, but more will undoubtedly be developed. To stack it with other MCPs with a few lines of code (with_tool(<your_tool_instance>)) or a few CLIs, maybe you might be interested in adapting your MCP to support the design requirements / Create a MCPStack MCP Tool wrapper around your MCP to offer other MCPs the chance to work with yours.
I hope it is useful! Feel free to ⭐ & share, it really helps!
Repo: https://github.com/MCP-Pipeline/MCPStack
Documentation: https://mcpstack.readthedocs.io/en/latest/
Cheers!