MCP Server

Connetra MCP Server

The Connetra MCP Server enables AI assistants to securely connect to your Connetra workspace using the Model Context Protocol (MCP). Once connected, supported AI clients can access your organization's metadata, data catalog, lineage, glossary, and other business context directly from Connetra.

What is MCP?

The Model Context Protocol (MCP) is an open standard that enables AI applications to communicate with external tools and services.

Instead of relying solely on the information available in a conversation, MCP allows AI assistants to securely retrieve live information from connected systems.

By connecting Connetra through MCP, your AI assistant can understand your organization's data landscape and answer questions using your Connetra workspace.

Why Use Connetra MCP?

With Connetra MCP connected, you can:

Search datasets and tables
Discover business glossary terms
Explore metadata and documentation
View data lineage
Find dashboards and related assets
Search schemas and columns
Understand data relationships
Ask natural language questions about your data catalog

Prerequisites

Before configuring Connetra MCP, ensure you have:

A Connetra workspace
A valid Connetra Access Token
Node.js installed
An MCP-compatible AI client

Get Your Connetra Access Token

Before connecting an AI client to Connetra, you need an MCP Client API Token. This token is used to authenticate your AI client and securely connect it to your Connetra workspace.

Note: Your AI client uses this token to access your Connetra workspace. Keep it secure and do not share it publicly.

Generate or Copy Your Access Token

Sign in to your Connetra account.
Click the Settings (⚙️) icon from the left navigation menu.
Select Workspace.
Under the Current Workspace section, locate the MCP Client API Token field.
Click Copy to copy the token to your clipboard.
Use this token when configuring your MCP client.

Common MCP Configuration

All supported clients use the same Connetra MCP configuration.

{
  "mcpServers": {
    "connetra": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "http://mcp.connetra.com/mcp",
        "--header",
        "Authorization: Bearer <YOUR_ACCESS_TOKEN>",
        "--allow-http"
      ],
      "env": {
        "MCP_TIMEOUT": "120000"
      }
    }
  }
}

Replace:

<YOUR_ACCESS_TOKEN> with your Connetra API token.

Setup Guides

Claude Desktop

Step 1 – Close Claude Desktop
Completely exit Claude Desktop before editing the configuration.

Step 2 – Open the Configuration File
Navigate to:
Settings → Developer → Edit Config
Claude will open the claude_desktop_config.json configuration file.

Step 3 – Configure Connetra MCP
Add the following configuration inside the mcpServers object.

{
  "mcpServers": {
    "connetra": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "http://mcp.connetra.com/mcp",
        "--header",
        "Authorization: Bearer <YOUR_ACCESS_TOKEN>",
        "--allow-http"
      ],
      "env": {
        "MCP_TIMEOUT": "120000"
      }
    }
  }
}

Step 4 – Save the Configuration
Save the file and ensure the JSON syntax is valid.

Step 5 – Restart Claude Desktop
Launch Claude Desktop again. The MCP server will start automatically.

Step 6 – Verify
Open a new chat. Click: ➕ → Add Connectors. Enable Connetra MCP, then begin chatting with your workspace.

Cursor

Step 1 – Open MCP Settings
Open Cursor. Navigate to: Settings → Tools & MCPs. Under Installed MCP Servers, click New MCP Server. Cursor opens the mcp.json configuration file.

Step 2 – Add Connetra MCP
Inside the mcpServers section, add:

"connetra": {
  "command": "npx",
  "args": [
    "mcp-remote",
    "http://mcp.connetra.com/mcp",
    "--header",
    "Authorization: Bearer <YOUR_ACCESS_TOKEN>",
    "--allow-http"
  ],
  "env": {
    "MCP_TIMEOUT": "120000"
  }
}

Step 3 – Save
Save the configuration file.

Step 4 – Open the Agent
Return to: Settings → Tools & MCPs. Click Meet the New Agents.

Step 5 – Enable MCP
Click the + icon in the chat. Select MCP Server. Enable Connetra if it is disabled. You are now ready to chat with your Connetra workspace.

Kiro

Step 1 – Open MCP Settings
Launch Kiro. Click the Kiro icon in the left sidebar. Locate the MCP Server section and click the Edit icon.

Step 2 – Open Workspace Configuration
Select the Workspace Config tab.

Step 3 – Add Configuration
Paste the following:

{
  "mcpServers": {
    "connetra": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "http://mcp.connetra.com/mcp",
        "--header",
        "Authorization: Bearer <YOUR_ACCESS_TOKEN>",
        "--allow-http"
      ],
      "env": {
        "MCP_TIMEOUT": "120000"
      }
    }
  }
}

Step 4 – Save
Save the configuration. The Connetra MCP server will be available immediately.

Antigravity

Step 1 – Open Settings
Launch Antigravity. Click Antigravity Settings in the bottom-right corner.

Step 2 – Open MCP Configuration
Navigate to: Customization → Installed MCP Servers → Open MCP Config. This opens the mcp_config.json file.

Step 3 – Add Connetra MCP
Add the following inside mcpServers.

"connetra": {
  "command": "npx",
  "args": [
    "mcp-remote",
    "http://mcp.connetra.com/mcp",
    "--header",
    "Authorization: Bearer <YOUR_ACCESS_TOKEN>",
    "--allow-http"
  ],
  "env": {
    "MCP_TIMEOUT": "120000"
  }
}

Step 4 – Save
Save the file. Restart Antigravity if required. The Connetra MCP server is now available.

Verify Your Connection

After completing the setup:

Open your AI client.
Start a new conversation.
Confirm that Connetra MCP appears in the list of available MCP servers.
Enable it if necessary.
Ask a question about your Connetra workspace.

For example:

Show me all Finance datasets.
Find the lineage for the Customer table.

If the AI returns information from your workspace, your MCP connection is working successfully.

Security

Connetra MCP authenticates every request using your Connetra Access Token.

Only authenticated users can access workspace information.
The AI assistant inherits the permissions of the authenticated user.
No additional access is granted beyond existing Connetra authorization policies.

Troubleshooting

If the MCP server does not appear:

Verify that Node.js is installed.
Confirm that your Access Token is valid.
Ensure the MCP endpoint is reachable.
Check that the JSON configuration is valid.
Restart the AI client after saving the configuration.
Verify that the Connetra MCP server is enabled in the client.

Integrations

Plans & Billing