Use TestKase with AI
Let Claude, Cursor or Copilot set up and run your tests on TestKase.
Your coding agent can connect your tests to TestKase for you. Give it the setup guide, and it knows how to use Automate, Test Management and Visual.
1. Teach your agent TestKase
Paste this into Claude, Cursor, Copilot or any coding agent. It works for Automate, Test Management and Visual. You can also read the guide yourself.
Read https://www.testkase.com/SKILL.md to set up TestKase testing.Automate runs Selenium and Playwright tests in browsers, Appium tests on iPhones and Android phones, Espresso tests on Android phones, and XCUITest tests on iPhones and iPads.
2. Let it manage your test cases (MCP)
With the MCP Server, your agent can create test cases, run cycles and pull reports in your TestKase projects.
In Claude Code, run this once in your terminal. Put your key in place of <key> yourself. Never paste it into the chat.
claude mcp add testkase -e TESTKASE_API_BASE_URL=https://api.testkase.com -e TESTKASE_PAT_TOKEN=<key> -- npx -y @testkase/test-management-mcpClaude Desktop, Cursor and GitHub Copilot use the same npx command in their MCP config. See MCP Server: Installation & Setup for each one.
3. Let it drive phones and browsers (MCP)
With the Device Farm MCP server, your agent can open a real phone or browser, tap and type on it, and explain why an Automate run failed.
In Claude Code, run this once in your terminal. Put your key in place of <key> yourself. Never paste it into the chat.
claude mcp add testkase-device-farm -e TESTKASE_PAT_TOKEN=<key> -- npx -y @testkase/device-farm-mcp4. Your key
One key for everything. Make it at Profile → API Keys (testkase.com/api-keys). It starts with xyz_. Keep it in an environment variable or CI secret, not in your code. See Managing API Keys.