Spark
Sign In
MCP Connector

Query OpenNeuro Neuroimaging Data via GraphQL

Query and explore OpenNeuro neuroimaging datasets via GraphQL directly from Claude Desktop or any MCP client.

Get from source
Works with githubcloudflare
QuentinCody
Own this? Claim it
88
Spark score
out of 100
Updated May 2025
Fresher alternatives ↓
Source checked Sep 15, 2026
Version 1.0.0
Models
universal
Add to Favorites

Why it matters

Access and query the OpenNeuro neuroimaging dataset API using GraphQL. Retrieve MRI, MEG, EEG, iEEG, and ECoG data for research and analysis.

Outcomes

What it gets done

01

Execute GraphQL queries against the OpenNeuro API.

02

Discover available fields and operations through schema introspection.

03

Query neuroimaging datasets, snapshots, and file listings.

04

Access public OpenNeuro data without requiring authentication.

Source

Get it from source

Spark does not host a copy of it.

Open source

Reports

Agent outcome reports

No reports yet

Capabilities

Tools your agent gets

openneuro_graphql_query

Execute GraphQL queries against the OpenNeuro API to access neuroimaging datasets

Overview

OpenNeuro MCP Server

OpenNeuro MCP Server wraps the OpenNeuro GraphQL API in a single MCP tool, openneuro_graphql, so an AI agent can list, search, and inspect public neuroimaging datasets. It supports arbitrary GraphQL queries with variables, covering dataset listing, lookup by ID, modality search, and summary metadata. Use it when you need to browse or query OpenNeuro's neuroimaging dataset catalog from an AI agent. It does not download or process the imaging files themselves.

What it does

OpenNeuro MCP Server exposes the OpenNeuro GraphQL API as a Model Context Protocol tool, letting an AI agent query and inspect neuroimaging datasets hosted on OpenNeuro directly from a conversation. It wraps the API in a single flexible tool, openneuro_graphql, that accepts any GraphQL query and variables, so an agent can list datasets, fetch a specific dataset by its OpenNeuro ID, filter datasets by modality, or pull a dataset's metadata and summary statistics without leaving the chat.

When to use - and when NOT to

Use it when you need to look up OpenNeuro datasets programmatically - browsing what's available, checking a dataset's subjects, sessions, tasks, modalities and file counts, or building a query against a known dataset ID such as ds000001. It is not a substitute for downloading or processing the imaging data itself: the server only talks to OpenNeuro's GraphQL endpoint and returns metadata, it does not fetch or manipulate the underlying neuroimaging files.

Capabilities

The server ships one general-purpose tool, openneuro_graphql, which runs arbitrary GraphQL queries (with variables) against the OpenNeuro API. Documented example queries cover: listing the first N datasets with id, name and created date; fetching a single dataset's public status, creation date, metadata (including modalities) and latest snapshot tag; searching datasets by modality such as MRI; and pulling a dataset's latest snapshot summary - subjects, sessions, tasks, modalities, total files and size. Because the tool is a thin GraphQL passthrough, any query the OpenNeuro API supports can be issued through it, not just the documented examples. A parameterized query pattern is also shown, using a variables object alongside a named GraphQL query such as GetDataset($id: ID!), for queries that need dynamic input.

How to install

The project runs as a local Python MCP server.

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install -r requirements.txt

Run it with python mcp_openneuro.py, then point an MCP client at that script - for Claude Desktop, add it to claude_desktop_config.json with the venv's Python interpreter as the command and the script path as the argument. Several test scripts are included (test_query.py, test_dataset_query.py, test_search_query.py, test_datasets_with_modality.py) to verify the connection and demonstrate each query pattern before wiring it into a client. Extending the server means editing mcp_openneuro.py and adding new methods decorated with @mcp.tool().

Who it's for

Researchers and developers working with open neuroimaging data who want an AI agent to browse or query OpenNeuro datasets conversationally, instead of hand-writing GraphQL calls or clicking through the OpenNeuro web interface. It is released under the MIT License with an added academic citation requirement: commercial and non-academic use follows standard MIT terms, while any academic publication built on results from this software must cite the project per its CITATION.md.

Source README

OpenNeuro MCP Server

Overview

This server provides MCP tools to interact with the OpenNeuro GraphQL API, allowing AI agents to easily access and manipulate neuroimaging datasets on OpenNeuro.

Installation

# Create a virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

Usage

Run the server:

python mcp_openneuro.py

Claude Desktop Configuration

To use this MCP server with Claude Desktop, add the following to your claude_desktop_config.json file (located at ~/Library/Application Support/Claude/claude_desktop_config.json on macOS):

{
  "mcpServers": {
    "openneuro": {
      "command": "/path/to/your/venv/bin/python",
      "args": [
        "/path/to/your/openneuro-mcp-server/mcp_openneuro.py"
      ],
      "options": {
        "cwd": "/path/to/your/openneuro-mcp-server"
      }
    }
  }
}

Replace /path/to/your/ with the actual path to your installation.

Available Tools

openneuro_graphql

Execute any GraphQL query against the OpenNeuro API.

Example usage in your AI agent's code:

result = await mcp.invoke("openneuro_graphql", {
    "query": "{ datasets(first: 5) { edges { node { id name } } } }"
})

Example Queries

Listing Datasets

{
  datasets(first: 10) {
    edges {
      node {
        id
        name
        created
      }
    }
  }
}

Getting a Specific Dataset

{
  dataset(id: "ds000001") {
    id
    name
    public
    created
    metadata {
      modalities
    }
    latestSnapshot {
      tag
      created
    }
  }
}

Search by Modality

{
  datasets(first: 5, modality: "MRI") {
    edges {
      node {
        id
        name
        created
      }
    }
  }
}

Dataset Metadata and Summary

{
  dataset(id: "ds000001") {
    id
    name
    metadata {
      modalities
      dataProcessed
      species
    }
    latestSnapshot {
      summary {
        subjects
        sessions
        tasks
        modalities
        totalFiles
        size
      }
    }
  }
}

Using Variables

Some queries require variables. Here's an example:

query = """
query GetDataset($id: ID!) {
  dataset(id: $id) {
    id
    name
    public
  }
}
"""

variables = {
    "id": "ds000001"
}

result = await mcp.invoke("openneuro_graphql", {
    "query": query,
    "variables": variables
})

Testing

Several test scripts are included to demonstrate usage:

  • test_query.py - Basic listing of datasets
  • test_dataset_query.py - Detailed information about a specific dataset
  • test_search_query.py - Search functionality example
  • test_datasets_with_modality.py - Filter datasets by modality

Run any test script:

python test_query.py

Development

To add more tools or functionality, modify the mcp_openneuro.py file and add new methods with the @mcp.tool() decorator.

FAQ

Common questions

Discussion

Questions & comments · 0

Sign In Sign in to leave a comment.