Skip to main content
Now that you have data in your Catalog, you can connect an AI agent to Nekt so it can work with that data on your behalf — no SQL or dashboards required. Nekt’s MCP Server gives agents like Claude, Cursor, ChatGPT, and any MCP-compatible client native access to your data through natural language. Once connected, your agent can:
  • Explore your data — discover tables, preview their contents, and understand schemas.
  • Query in plain English — ask questions and let the agent generate and run SQL against your warehouse.
  • Read files — browse layers, volumes, and files stored in Nekt.
  • Build and run pipelines — create Queries and Notebooks, schedule them, and trigger runs.
  • Learn about Nekt — ask about platform features and capabilities without switching context.
The MCP Server is available on all paid plans.

​
Enable the MCP Server

In order to connect an agent, first enable the server and create a token:
  1. Go to Integrations → MCP Server and click Configure.
  2. Click Enable MCP Server. Setup varies slightly by cloud — see the setup guide for GCP and AWS specifics.
  3. In the MCP Tokens section, create a token. Each token can be scoped to specific tables, so you control exactly what data your agent can see.
If you connect via OAuth and don’t have a token yet, one is created automatically with the same permissions as your user account.

​
Connect your AI agent

The MCP Server supports two authentication methods, depending on how your agent connects:
  • OAuth 2.1 — for desktop apps like Claude Desktop and Cursor. Point the app to your MCP server URL and log in with your Nekt credentials.
  • Bearer token — for CLI tools and automation platforms like N8N or Make. Add the server to your MCP client configuration:
{
  "mcpServers": {
    "Nekt": {
      "type": "http",
      "url": "<MCP_ENDPOINT>",
      "headers": {
        "Authorization": "Bearer <MCP_TOKEN>"
      }
    }
  }
}
Replace <MCP_ENDPOINT> with the endpoint from the MCP Server page in Nekt, and <MCP_TOKEN> with the token you created. See Authentication for full details.

​
Start asking

With the agent connected, ask questions in natural language and let it pick the right tools:
“Which tables should I use to analyze customer churn?”
“How many leads were created each month?”
“Create a daily Query that aggregates revenue by customer into the trusted layer.”

​
Connect a specific tool

Step-by-step guides for the most common AI agents:

Claude

Connect Claude to your data in Nekt.

ChatGPT

Connect ChatGPT using OAuth authentication.

Cursor & others

Connect any MCP-compatible client.
Nice! Your AI agent can now work directly with your data in Nekt 🤖 ➡️ Next, let’s explore your data with Nekt’s built-in AI assistant.
Need Help? If you encounter any issues during onboarding or need assistance, feel free to reach out to our support team. We’re here to help you get started.