@brain-bbqs/neuro-mcp

NeuroMCP

GlobalModelContextProtocol using FastMCP for Neuroscience Python Packages

NeuroMCP is a scalable, Pythonic framework for integrating neuroscience Python packages through the Model Context Protocol (MCP) using FastMCP. It provides a standardized way to expose neuroscience tools, ontologies, and neural device interfaces for agentic AI workflows.

Architecture

neuromcp/
├── src/neuromcp/
│   ├── __init__.py           # Package entry point
│   ├── server.py             # Main MCP server implementation
│   ├── registry.py           # Provider registry and discovery
│   ├── cli.py                # Command-line interface
│   ├── config/
│   │   ├── __init__.py
│   │   └── settings.py       # Configuration management
│   ├── providers/            # Provider implementations
│   │   ├── __init__.py
│   │   ├── base.py                   # Base provider interface
│   │   ├── ontology_provider.py      # Ontology & BIDS validation
│   │   └── neural_device_provider.py # Device translation
│   └── utils/                # Utility functions
│       ├── __init__.py
│       └── validation.py     # Data validation utilities
├── tests/
│   ├── unit/                 # Unit tests
│   └── integration/          # Integration tests
├── examples/                 # Usage examples
├── pyproject.toml           # Project configuration
└── README.md                # This file

Installation

Using pip (when published)

pip install neuromcp

Quick Start

Using the CLI

# List available providers
neuromcp list-providers

# Run the server with auto-discovery
neuromcp run

# Run with custom configuration
neuromcp run --config config.yaml

# Run with specific options
neuromcp run --log-level DEBUG --auto-discover true

Built-in Providers

Ontology Provider

The Ontology Provider offers comprehensive support for neuroscience standards and ontologies through specialized subproviders:

BIDS (Brain Imaging Data Structure)

• Brain Imaging Data Structure validation
• Quality control for neuroimaging datasets
• Upload readiness checks for OpenNeuro and other databases
• Comprehensive dataset structure analysis
• Example prompts for scientists
• See Quality Control Guide for details

NWB (Neurodata Without Borders)

• NWB format information and best practices
• Data structure validation
• File organization guidance
• Integration examples for electrophysiology, imaging, and behavior data

NBO (Neuro Behavior Ontology)

• Behavioral term search and validation
• Semantic annotation for behavioral experiments
• Ontology-driven data validation
• Support for owlready2 and rdflib

HED (Hierarchical Event Descriptors)

• Event annotation and validation
• HED tag string validation
• Schema version management
• Best practices for experimental event tagging

General Ontology Tools

• Neuroscience term standardization
• Ontology validation
• Format translation (abbreviated ↔ standard)

Neural Device Provider

• Device specifications and metadata
• Data format translation
• Recording parameter validation

License

MIT License - see LICENSE file for details