Limetest

Limetest is the most light weight end to end testing framework with AI capabilities that can run in your CI workflows. Define your test cases in natural language and let AI handle the execution.

Key Features

• Optimized for AI: Define your test cases in plain language and let AI execute them end to end.

• Lightweight & Efficient: Leverages Playwright snapshot instead of pixel analysis for faster, more reliable execution.

• Vision Capabilities: Falls back to vision mode when snapshot mode fails during more sophisticated test scenarios.

Installation

npm install @limetest/limetest

npx playwright install

User data directory

limtest will launch Chrome browser with the new profile, located at

- `%USERPROFILE%\AppData\Local\ms-limetest\mcp-chrome-profile` on Windows
- `~/Library/Caches/ms-limetest/mcp-chrome-profile` on macOS
- `~/.cache/ms-limetest/mcp-chrome-profile` on Linux

Usage

Run Tests

Use --headless for running tests headlessly in CI workflows

npx limetest example

limetest MCP Server

https://github.com/user-attachments/assets/b801f239-dc66-4b3b-bcf2-42e2a9a68721

A Model Context Protocol (MCP) server powered by Playwright that streamlinse end to end testing for your MCP client.

Use Cases

• Automated testing planned and executed by LLMs

Example config

After cloning this repo, build and add the E2E MCP server to your MCP Client as such: Notice that you need OpenAI API key to run this MCP server in end to end mode.

npm install @limetest/mcp

npx playwright install

Then:

{
    "mcpServers": {
        "limetest": {
            "command": "npx",
            "args": [
                "npx @limetest/mcp",
                "--api-key=<your openai api key>"
            ]
        }
    }
}

All the logged in information will be stored in that profile, you can delete it between sessions if you'd like to clear the offline state.

Acknowledgements

Limetest is based on Microsoft's Playwright MCP and optimized for automated end-to-end testing as a standalone framework. This project is distributed under the Apache 2.0 License.