alumnium-hq/alumnium

What it solves

Alumnium simplifies end-to-end (E2E) testing by replacing complex, fragile test scripts with AI-driven interactions. It reduces the need for manual selector management and rigid assertions by allowing developers to describe test actions and verifications in plain language.

How it works

It acts as an AI-native library and Model Context Protocol (MCP) server that integrates with existing test automation tools like Playwright, Selenium, Appium, and Maestro. Instead of writing code to find specific HTML elements, users can use commands like .do() or .act() to perform actions (e.g., "type 'selenium' into the search field") and .check() to verify results (e.g., "page title contains selenium").

Who it’s for

QA engineers and developers who want to make their E2E tests more robust and easier to maintain using AI, as well as users of AI agents (like Claude Code) who want to automate application testing via MCP.

Highlights

• Multi-framework support: Works with Selenium, Playwright, Appium, and Maestro.
• Multi-language support: Available as a library for Java, Python, and TypeScript.
• MCP Integration: Can be added to AI agents to allow them to test applications directly.
• Natural language assertions: Verifies application state using plain-text descriptions.