Skip to main content
Testkube 2.10.0 is out! Granular Metrics, AI MCP improvements, Organization management, and much more. Read More

Testkube MCP Server Configuration Examples

This guide provides configuration examples for integrating the Testkube MCP Server with various AI tools and development environments.

tip

Before configuring your AI tools, make sure you've set up the Testkube MCP Server using one of these methods:

GitHub Copilot (VS Code)

GitHub Copilot with Agent mode in VS Code provides excellent agentic capabilities for multi-step testing workflows.

  1. Create or edit your MCP configuration file:

    Location: ~/Library/Application Support/Code/User/mcp.json (macOS) or %APPDATA%\Code\User\mcp.json (Windows)

    {
    "servers": {
    "testkube": {
    "type": "http",
    "url": "https://api.testkube.io/organizations/tkcorg_YOUR_ORG_ID/environments/tkcenv_YOUR_ENV_ID/mcp"
    }
    }
    }

    Replace tkcorg_YOUR_ORG_ID and tkcenv_YOUR_ENV_ID with your actual values. No API token needed — VS Code handles OAuth authentication automatically via a browser login.

    Learn more about the hosted endpoint →

Using the Hosted Endpoint with API Token

If you prefer using an API token (e.g., for shared or automated setups):

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

Replace tkcorg_YOUR_ORG_ID, tkcenv_YOUR_ENV_ID, and YOUR_API_TOKEN_HERE with your actual values.

  1. Restart VS Code to load the new configuration

  2. Test the integration:

    • Open GitHub Copilot Agent mode
    • Try a prompt like: list my testkube workflows

Using the CLI

  1. Create or edit your MCP configuration file:

    Location: ~/Library/Application Support/Code/User/mcp.json (macOS) or %APPDATA%\Code\User\mcp.json (Windows)

    {
    "servers": {
    "testkube": {
    "command": "testkube",
    "args": ["mcp", "serve"],
    "type": "stdio"
    }
    },
    "inputs": []
    }

    Learn more about CLI setup →

  2. Restart VS Code to load the new configuration

  3. Test the integration:

    • Open GitHub Copilot Agent mode
    • Try a prompt like: list my testkube workflows

With GitHub Copilot Agent mode, you can perform complex multi-step operations:

Help me debug the last failed execution of my api-tests workflow.
Show me the logs, check for any artifacts, and suggest what might be wrong.
Best Experience

We recommend using Claude Sonnet 4 with GitHub Copilot for the most sophisticated reasoning capabilities.

Cursor

Cursor provides powerful AI assistance with built-in MCP support.

Add the hosted endpoint configuration to your Cursor MCP settings:

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

Replace tkcorg_YOUR_ORG_ID, tkcenv_YOUR_ENV_ID, and YOUR_API_TOKEN_HERE with your actual values.

Learn more about the hosted endpoint →

Using the CLI

Add Testkube server configuration to your Cursor MCP settings:

{
"mcpServers": {
"testkube": {
"command": "testkube",
"args": ["mcp", "serve"],
"type": "stdio"
}
}
}

Learn more about CLI setup →

Usage Examples

Create a test workflow that checks my REST API endpoints, then run it and show me the results.
Look at my recent test failures and help me understand what's causing them. Check the logs and artifacts.

Claude

Claude Desktop

For direct interaction with Claude through the desktop application.

Using the Hosted Endpoint with OAuth

You can add the Testkube MCP server via Settings → Connectors → Add custom connector in Claude Desktop. Paste the endpoint URL and Claude Desktop handles OAuth authentication automatically — you'll see a browser popup to log in.

info

Claude Desktop's claude_desktop_config.json only supports local stdio servers (command + args). To use the hosted endpoint with OAuth, add it as a custom connector via Settings → Connectors, or use the CLI/Docker options below.

  1. Create or edit Claude Desktop config:

    Location: ~/Library/Application Support/Claude/claude_desktop_config.json (macOS)

    {
    "mcpServers": {
    "testkube": {
    "command": "testkube",
    "args": ["mcp", "serve"]
    }
    }
    }

    Learn more about CLI setup →

  2. Restart Claude Desktop

Using Docker

  1. Create or edit Claude Desktop config:

    Location: ~/Library/Application Support/Claude/claude_desktop_config.json (macOS)

    {
    "mcpServers": {
    "testkube": {
    "command": "docker",
    "args": [
    "run", "--rm", "-i",
    "-e", "TK_ACCESS_TOKEN=YOUR_API_TOKEN_HERE",
    "-e", "TK_ORG_ID=tkcorg_YOUR_ORG_ID",
    "-e", "TK_ENV_ID=tkcenv_YOUR_ENV_ID",
    "kubeshop/testkube-mcp-server:latest", "mcp", "serve"
    ]
    }
    }
    }

    Replace tkcorg_YOUR_ORG_ID, tkcenv_YOUR_ENV_ID, and YOUR_API_TOKEN_HERE with your actual values.

    Learn more about Docker setup →

  2. Restart Claude Desktop

Claude Code

claude mcp add --transport http testkube https://api.testkube.io/organizations/tkcorg_YOUR_ORG_ID/environments/tkcenv_YOUR_ENV_ID/mcp

Claude Code handles OAuth automatically — a browser window opens for you to log in. No tokens needed.

Learn more about the hosted endpoint →

Using Docker

claude mcp add testkube -- docker run --rm -i \
-e TK_ACCESS_TOKEN=${TK_ACCESS_TOKEN} \
-e TK_ORG_ID=${TK_ORG_ID} \
-e TK_ENV_ID=${TK_ENV_ID} \
kubeshop/testkube-mcp-server:latest mcp serve

Learn more about Docker setup →

Usage

Claude Desktop provides conversational access to your Testkube resources:

Can you show me all my test workflows and their recent execution status?

I need to create a new workflow for testing my Python application.
Can you help me set this up?

Example Claude Workflows

Automated Debugging Session

I need help with my failed test execution "api-tests-123".
Can you:
1. Get the execution details
2. Show me the logs
3. Check if there are any artifacts
4. Analyze what went wrong
5. Suggest next steps

Test Analysis and Reporting

Analyze my test execution trends for the last week.
Show me which workflows are failing most often and help me create a summary report for my team.