Knowledge Graph MCP Server

Connect your AI assistant to the EBRAINS Knowledge Graph — search neuroscience datasets, models, and software through the Model Context Protocol.

What is this?

This is a remote MCP server that gives AI assistants (like Claude, ChatGPT, Gemini) direct access to the EBRAINS Knowledge Graph, a metadata store for neuroscience research data, models, and software hosted on the EBRAINS research infrastructure.

Once connected, you can ask your AI assistant to search for datasets, look up model details, explore controlled vocabularies, and more — all without leaving the conversation.

Connect your AI assistant

Claude Desktop & claude.ai

Works with the Claude Desktop app and claude.ai on the web.

Open Settings → Connectors

Click the menu icon, go to Settings, and select Connectors.

Add this server URL

Click Add custom connector, choose a name, and paste the following into the URL field:

https://fairgraph-mcp.apps.dev-adacloud.ebrains.eu/mcp

Log in with your EBRAINS account

Click Connect. A browser window will open for you to authenticate with your EBRAINS account. Once logged in, the connection is established automatically.

Claude Code (CLI)

Add the remote server to your project's .mcp.json:

{
  "mcpServers": {
    "ebrains-kg": {
      "type": "url",
      "url": "https://fairgraph-mcp.apps.dev-adacloud.ebrains.eu/mcp"
    }
  }
}

Claude Code will open a browser for OAuth authentication on first use.

Works with the ChatGPT Desktop app (Pro, Team, Enterprise, or Edu plan required).

Enable Developer Mode

Go to Settings → Apps & Connectors → Advanced settings and enable Developer Mode.

Create a new connector

Back in Apps & Connectors, click Create, enter a name (e.g. "EBRAINS Knowledge Graph"), and paste the URL:

https://fairgraph-mcp.apps.dev-adacloud.ebrains.eu/mcp

Start a chat

Create a new chat and select the connector as a source. You will be prompted to log in with your EBRAINS account.

Works with VS Code using GitHub Copilot.

Add to .vscode/mcp.json in your workspace, or open the Command Palette and run MCP: Open User Configuration for global setup:

{
  "servers": {
    "ebrains-kg": {
      "type": "http",
      "url": "https://fairgraph-mcp.apps.dev-adacloud.ebrains.eu/mcp"
    }
  }
}

VS Code will handle the OAuth flow automatically when the server is first used.

In Cursor, go to Settings → Developer → Edit Config and add:

{
  "mcpServers": {
    "ebrains-kg": {
      "url": "https://fairgraph-mcp.apps.dev-adacloud.ebrains.eu/mcp"
    }
  }
}

Cursor will prompt you to authenticate with your EBRAINS account on first use.

Works with the Gemini CLI. Add to your settings.json:

{
  "mcpServers": {
    "ebrains-kg": {
      "httpUrl": "https://fairgraph-mcp.apps.dev-adacloud.ebrains.eu/mcp"
    }
  }
}

The CLI will handle the OAuth flow automatically. Run /mcp in the CLI to check connection status.

This server works with any MCP-compatible client that supports remote servers with OAuth authentication. Use the following endpoint:

https://fairgraph-mcp.apps.dev-adacloud.ebrains.eu/mcp

For local installation (stdio transport), see the project repository.

What can you do with it?

  • Search for datasets, models, software, and other research products in the Knowledge Graph
  • Get details on specific items, including their metadata, authors, and linked resources
  • Explore vocabularies of controlled terms used in neuroscience (species, brain regions, techniques, etc.)
  • Compare multiple items side by side in a single request

Example queries

Once connected, try asking your AI assistant:

  • "Search for datasets about the hippocampus"
  • "Find models by Migliore"
  • "What species are available as controlled terms?"
  • "Get details on dataset version <uuid>"