Skip to main content
Claude is a powerful and versatile AI tool, but it becomes even stronger when connected to your data. You can connect Claude to Nekt through Claude.ai (web) or Claude Code (CLI).
Requires a paid Nekt plan (for MCP Server access).

​
1. Activate your MCP Server

The connection is made by activating the MCP Server in your workspace. Check this guide for instructions.

​
2. Connect Claude to Nekt

Choose the setup that matches how you use Claude.

Claude.ai (Web & Desktop)

Check the video below for a visual guide:

​
Add the MCP as a custom connector

Access Claude, and click on the + icon on the prompt box to add a connector.In the menu, select Connectors and then Manage Connectors.A new screen will open. Click the + icon and select Add custom connector. Fill in the required information, which you’ll find in the Connection fields section inside Nekt.
  1. Name: we suggest “Nekt”
  2. Remote MCP Server URL: paste the value of the Endpoint field
  3. Open the advanced settings section and paste the values of the OAuth Client ID and OAuth Client Secret fields.
Click on Add.

​
Connect to Nekt

Now you should have a new item called “Nekt” (or the name you previously defined) in the list of Not Connected connectors.Click on it and then in the Connect button in the middle of the screen.Log in with your Nekt credentials and a list of your tokens will be displayed.

​
Select your tokens

Tokens authenticate your requests to Nekt. They give Claude access to the tables associated with each token.
  • If you haven’t created any custom tokens yet, you’ll see a default one created automatically with Full Access. This token accesses all tables in your workspace.
  • If you already have a custom one created, you’ll see them all on the list. Select the one you want to use and click on Select Token to advance.
Tokens created during the OAuth flow use the creating user’s permissions by default. Tokens created manually can be scoped to specific tables for fine-grained access control.
You’ll see a success message and you’ll be redirected to Claude.app.

​
Manage tool permissions

Nekt MCP Server provides nine tools that AI agents can use to interact with your data. You can learn more about them here.Make sure to grant the Always allow permission to these tools.

​
Get your endpoint and token

Go to Integrations > MCP Server in the Nekt platform. You’ll need two values:
  • Endpoint — the MCP server URL displayed at the top of the page
  • Token — a bearer token created in the MCP Tokens section
If you don’t have a token yet, create one. You can scope it to specific tables or grant full access.

​
Add the MCP server to your configuration

Open your .claude.json file (in your project root or home directory) and add the following block:
{
  "mcpServers": {
    "Nekt": {
      "type": "http",
      "url": "MCP_ENDPOINT",
      "headers": {
        "Authorization": "Bearer MCP_TOKEN"
      }
    }
  }
}
Replace MCP_ENDPOINT with the endpoint from the MCP Server page and MCP_TOKEN with the token you created.

​
3. Explore your data

Once connected, ask Claude a question about your data. Here’s an example: 
Using /nekt give me insight about your_table_name table.
While it processes your request, you can see the steps taken and the queries executed to get to a final answer. It may take a few minutes.
We’d love to know what you’re building with Nekt! Share your projects with us through any of our communication channels.

​
Privacy & Security

Your data privacy is a core principle of the Nekt platform. When you connect Claude to Nekt through the MCP Server, Claude does not receive direct database credentials or unrestricted access to your infrastructure. Instead, Claude interacts with your data exclusively through the secure tools exposed by the Nekt MCP Server, which act as a controlled gateway to the Nekt API. These tools enforce the same authentication, authorization, and permission rules used throughout the Nekt platform. Access is scoped using tokens and user permissions, meaning Claude can only access the tables and actions explicitly allowed by the token you selected during connection. This architecture ensures that all queries, analysis, and data interactions remain governed, auditable, and restricted to your workspace permissions. When using Claude (via Claude.ai or Claude Code), your interactions are subject to Anthropic’s privacy settings and policies. Anthropic allows users to control whether their conversations may be used for model training through their Privacy Settings. We recommend reviewing the official privacy documentation from the respective providers to understand their data handling policies and selecting the configuration that best aligns with your organization’s security requirements.

​
Need help?

Contact our support team if you encounter issues during setup.