Skip to main content
Testkube 1.21 is out! A new Execution Dashboard, Webhook Credentials and Test/Suite Deprecation are some of the major enhancements, Read More

Testkube Control Plane MCP Endpoint

The Testkube Control Plane provides an MCP endpoint that allows you to connect directly to Testkube without running a local MCP server. This is the recommended approach for most users.

tip

This is the easiest way to get started with the Testkube MCP Server - no local installation required!

Why Use the Control Plane MCP Endpoint?

  • No local setup - No need to install the Testkube CLI or Docker
  • Remote access - Connect to Testkube from any location without local setup
  • Always available - The endpoint is always running and accessible
  • Simplified deployment - No need to manage a local MCP server process
  • Centralized configuration - Manage authentication and permissions through the Control Plane
  • Team collaboration - Multiple users can connect to the same endpoint
  • Automatic updates - Control Plane updates are handled automatically
  • Consistent performance - Leverages the Control Plane's infrastructure

Prerequisites

  • Access to a Testkube Environment - You need an active Testkube organization and environment
  • API Token - A valid Testkube API token with appropriate permissions
  • An AI tool that supports MCP with SSE transport - Such as Claude Desktop, Cursor, VS Code with GitHub Copilot, or custom MCP clients

Endpoint URL Structure

The Control Plane MCP endpoint follows this URL pattern:

https://api.testkube.io/organizations/{organization_id}/environments/{environment_id}/mcp

Replace:

  • {organization_id} with your Testkube organization ID (e.g., tkcorg_076487a004a7f6fb)
  • {environment_id} with your environment ID (e.g., tkcenv_d19e797ff2c1449b)
Finding Your IDs

You can find your organization and environment IDs in the Testkube Dashboard URL or by running:

testkube get context

Configuration Example

To use the Control Plane MCP endpoint with an AI tool that supports SSE (Server-Sent Events) transport:

{
"mcpServers": {
"testkube": {
"url": "https://api.testkube.io/organizations/tkcorg_076487a004a7f6fb/environments/tkcenv_d19e797ff2c1449b/mcp",
"transport": {
"type": "sse"
},
"headers": {
"Authorization": "Bearer YOUR_API_TOKEN_HERE"
}
}
}
}

Replace:

  • tkcorg_076487a004a7f6fb with your actual organization ID
  • tkcenv_d19e797ff2c1449b with your actual environment ID
  • YOUR_API_TOKEN_HERE with your Testkube API token

Obtaining an API Token

  1. Log in to your Testkube Dashboard
  2. Navigate to Organization SettingsAPI Tokens
  3. Create a new API token with appropriate permissions
  4. Copy the generated API token (you won't be able to see it again)

See API Token Management for more details.

Control Plane Endpoint in the Testkube Dashboard

The Testkube Dashboard provides convenient access to the Control Plane MCP endpoint for your Testkube Environments together with configuration examples for common IDEs, simply navigate to the Environment Settings page and select the MCP Server option:

MCP Server Settings

Testkube Cloud vs Self-Hosted

The MCP endpoint feature shown in the screenshots above is enabled by default in Testkube Cloud.

For self-hosted installations, you must enable this feature in your Helm configuration (see below) before the MCP Server settings tab and endpoint become available.

Enabling MCP for Self-Hosted Deployments

Self-Hosted Configuration Required

For self-hosted Testkube Enterprise installations, the MCP endpoint is disabled by default. You must enable it via Helm configuration before it becomes available.

Step 1: Enable MCP in Helm Configuration

Add the following environment variables to your Helm values file:

testkube-cloud-api:
additionalEnvVars:
- name: MCP_ENABLED
value: "true"

testkube-cloud-ui:
additionalEnv:
MCP_ENABLED: true

Step 2: Enable MCP per Environment

After deploying with MCP enabled, you need to enable it for each environment:

  1. Log in to your Testkube Dashboard
  2. Navigate to Environment SettingsMCP Server
  3. Toggle the MCP Endpoint switch to "Enabled"
  4. Copy the configuration for your preferred IDE

Step 3: Configure Your AI Tool

For self-hosted instances, use your custom control plane URL in the MCP configuration:

{
"mcpServers": {
"testkube": {
"url": "https://your-control-plane.example.com/organizations/{organization_id}/environments/{environment_id}/mcp",
"transport": {
"type": "sse"
},
"headers": {
"Authorization": "Bearer YOUR_API_TOKEN_HERE"
}
}
}
}

Replace:

  • your-control-plane.example.com with your self-hosted control plane domain
  • {organization_id} with your organization ID
  • {environment_id} with your environment ID
  • YOUR_API_TOKEN_HERE with your API token