MCP Bio‑Modelling Servers

<!-- mcp-name: io.github.marcorusc/NeKo --> <!-- mcp-name: io.github.marcorusc/MaBoSS --> <!-- mcp-name: io.github.marcorusc/PhysiCell -->

This repository centralizes Model Context Protocol (MCP) servers that wrap Python‑based mechanistic / systems biology modelling tools. Each subfolder contains a server.py entrypoint plus a README describing the specific tool interface.

Current servers (see their own READMEs & upstream docs):

All servers are Python processes speaking MCP over stdio.

Installation

Option A — pip (recommended)

pip install mcp-biomodelling-servers

Then run any server directly:

mcp-neko-server
mcp-maboss-server
mcp-physicell-server

Option B — uvx (no install needed)

uvx --from mcp-biomodelling-servers mcp-neko-server
uvx --from mcp-biomodelling-servers mcp-maboss-server
uvx --from mcp-biomodelling-servers mcp-physicell-server

Option C — from source (Conda, full control)

Clone this repo and set up a Conda environment with all dependencies (see Environment Assumption below).

MCP Background

The Model Context Protocol standardizes how external tools expose tools and resources to AI assistants / IDEs. Spec & introduction: https://modelcontextprotocol.io/docs/getting-started/intro

Each server.py advertises modelling actions (e.g. run simulations, manage sessions) to any MCP‑aware client (e.g. VS Code with GitHub Copilot Chat MCP support).

Repository Layout

MaBoSS/    # MaBoSS MCP server (Boolean / stochastic models)
NeKo/      # NeKo MCP server
PhysiCell/ # PhysiCell settings / sessions MCP server
README.md

Consult the README within each tool folder for: purpose, required Python packages, and any model/data file expectations. Installation instructions for the modelling tools themselves live there (or in the upstream project links above) — they are intentionally not duplicated here.

Environment Assumption

All tools are Python‑based. Create (and manage) a single Conda environment that contains the dependencies for MaBoSS, NeKo, and PhysiCell. The exact creation commands are up to you (not prescribed here). Once created, note the absolute path to its Python interpreter (e.g. /home/you/miniforge3/envs/mcp_modelling/bin/python).

Configure in VS Code (GitHub Copilot Chat / MCP)

1. Open VS Code and ensure the Copilot Chat (or other MCP-capable) extension is installed.
2. Press Ctrl + Shift + P → "MCP: Open Configuration" (or edit ~/.config/Code/User/mcp.json directly).
3. Add entries for each server.

Simple setup (uvx / pip install)

If you installed via pip or want to use uvx, no paths are needed:

{
  "servers": {
    "neko": {
      "type": "stdio",
      "command": "uvx",
      "args": ["--from", "mcp-biomodelling-servers", "mcp-neko-server"]
    },
    "maboss": {
      "type": "stdio",
      "command": "uvx",
      "args": ["--from", "mcp-biomodelling-servers", "mcp-maboss-server"]
    },
    "physicell": {
      "type": "stdio",
      "command": "uvx",
      "args": ["--from", "mcp-biomodelling-servers", "mcp-physicell-server"]
    }
  }
}

Advanced setup (Conda environment, from source)

Use this if you need a custom Conda environment (e.g. for native MaBoSS binaries or local development):

{
  "servers": {
    "maboss": {
      "type": "stdio",
      "command": "/home/you/miniforge3/envs/mcp_modelling/bin/python",
      "args": [
        "/absolute/path/to/mcp-biomodelling-servers/MaBoSS/server.py"
      ],
      "env": {
        "PATH": "/home/you/miniforge3/envs/mcp_modelling/bin:${Path}",
        "CONDA_PREFIX": "/home/you/miniforge3/envs/mcp_modelling"
      }
    },
    "neko": {
      "type": "stdio",
      "command": "/home/you/miniforge3/envs/mcp_modelling/bin/python",
      "args": [
        "/absolute/path/to/mcp-biomodelling-servers/NeKo/server.py"
      ]
    },
    "physicell": {
      "type": "stdio",
      "command": "/home/you/miniforge3/envs/mcp_modelling/bin/python",
      "args": [
        "/absolute/path/to/mcp-biomodelling-servers/PhysiCell/server.py"
      ]
    }
  }
}

Replace /home/you/... and /absolute/path/to/... with your actual directories. Keep all three servers referencing the same Conda interpreter to share installed libraries.

After saving, reload / restart VS Code so the MCP client reconnects.

Activation / usage guidance in VS Code: https://code.visualstudio.com/docs/copilot/chat/mcp-servers

You should then see the servers' tools listed in the Copilot Chat "/tools" (or similar) UI. Invoke them by name with required parameters.

Adding Another Server

1. Create a new folder with server.py and a README describing the underlying modelling tool and dependencies.
2. Follow existing server structure for registering MCP tools.
3. Update your mcp.json with a new block (use the same Conda Python path).
4. Document any additional env vars in that folder README.

License

Project is MIT (see existing LICENSE file). Underlying tools retain their own licenses — consult upstream repositories.

Quick Reference

Happy modelling!