NA-MIC Project Weeks

Back to Projects List

MHub.ai MCP server for Model Provision and Execution

Key Investigators

• Leonard Nürnberg (AIM Lab, USA)

Project Description

MCP servers have significantly shaped and improved the abilities of agentic AI systems. They provide AI agents with structured, on-demand capabilities (tools) and information (resources / promts). The aim is, to enable AI systems to find suitable AI models, answer questions about specific models and upon request run suitable AI models. The secondary aim is to enable automatic conversion of open source models into the MHub.ai deployment format.

In this project, we want to enable AI agents to interact with MHub.ai models by providing a (local) MCP server for MHub.ai. The public api, structured metadata and output-level information (e.g., anatomic structures segmented by each model) are already present and can be directly exposed through the server. The detailed technical documentation and step-by-step tutorials furthermore might already be a decent starting point to automatically generate MHub.ai project structure, Dockerfile, Module wrapper class, default workflow configuration and metadata json, to automate or speed-up the contribution process.

Objective

1. Local MCP server to interact with the MHub.ai model repository
2. Provision of tools and guidelines for automatic model deployment

If that works:

1. Implementation of 1-2 PoC models using the MPC.

Approach and Plan

1. Describe specific steps of what you plan to do to achieve the above described objectives.

Progress and Next Steps

1. Describe specific steps you have actually done.

Illustrations

No response

Background and References

No response