Nuclino supports the Model Context Protocol (MCP), an open standard that lets AI assistants access your workspace content directly. This means you can ask your AI assistant to search your documentation, create new pages, or update existing content. All through natural conversation.

What is MCP?

MCP is an open protocol that allows AI applications to securely connect to external data sources. When you connect Nuclino via MCP, your AI assistant can:

• Search your workspaces for relevant information.

• Read the full content of any page.

• Create new items and collections.

• Update existing pages.

• Organize content by moving items to trash.

Connecting Nuclino to an MCP Client

MCP is supported by a growing number of AI applications, including Claude Desktop, Cursor, Windsurf, and others. The setup process varies by application, but generally involves adding the Nuclino server URL to your MCP configuration.

Server Details

General Setup Steps

1. Open your AI application's MCP or integration settings.

2. Add a new MCP server with the URL https://api.nuclino.com/mcp.

3. Restart the application if required.

4. When you first interact with Nuclino through the AI, you'll be prompted to authorize access.

5. Sign in to your Nuclino account and approve the connection.

Example Configuration

Many MCP clients use a JSON configuration file. The Nuclino entry typically looks like this:

{
  "mcpServers": {
    "nuclino": {
      "url": "https://api.nuclino.com/mcp"
    }
  }
}

Refer to your AI application's documentation for the exact configuration format and file location.

Using an API Key (for clients without OAuth support)

If your MCP client doesn't support OAuth, you can authenticate using an API key instead. After creating an API key, you can configure your MCP client with the bearer token:

{
  "mcpServers": {
    "nuclino": {
      "url": "https://api.nuclino.com/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Replace YOUR_API_KEY with the key you created.

Available Capabilities

Once connected, your AI assistant can use these tools to interact with Nuclino:

Example Use Cases

Once Nuclino is connected, you can ask your AI assistant things like:

• "Search my Nuclino workspace for our API documentation"

• "Create a new page in Nuclino titled 'Meeting Notes' with a summary of what we discussed"

• "Find the onboarding checklist and add a new item for security training"

• "What does our documentation say about deployment procedures?"

Managing Your Connection

You can view and manage your MCP connections from your Nuclino profile settings under API Keys. Each MCP connection creates an API key labeled with the application name (e.g., "MCP: claude.ai" or "MCP: cursor.com").

To disconnect an AI assistant from your Nuclino account, revoke the API key.

Security & Privacy

• OAuth 2.0 with PKCE: Industry-standard secure authorization flow

• Scoped Access: The AI can only access workspaces and teams your account has permission to view or edit

• No Password Sharing: Your Nuclino password is never shared with third-party applications

• Revocable: You can disconnect any application at any time from your account settings

• Rate Limited: API access is rate-limited to prevent abuse

For Developers

If you're building an MCP client or integration:

• Endpoint: https://api.nuclino.com/mcp

• Protocol: MCP over HTTPS with JSON responses (stateless mode)

• Authentication: OAuth 2.0 with PKCE

• OAuth Metadata: https://api.nuclino.com/.well-known/oauth-authorization-server