This project provides an automated QA testing solution for websites using multiple browser agents. It targets developers and QA engineers looking to identify UI bugs, broken links, and accessibility issues efficiently. The benefit is a more robust and reliable web application through automated, comprehensive testing.

How It Works

The system utilizes an MCP (Multi-Agent Control Plane) server that orchestrates multiple Browser-Use agents. These agents navigate and interact with specified websites, performing automated checks. This approach allows for parallel testing, increasing coverage and speed compared to manual or single-agent testing.

Quick Start & Requirements

• Install dependencies: uv venv, source .venv/bin/activate, uv pip install -e .
• Add MCP server via CLI: claude mcp add vibetest /full/path/to/vibetest-use/.venv/bin/vibetest-mcp -e GOOGLE_API_KEY="your_api_key"
• Requirements: Python 3.11+, Google API key (Gemini 2.0 Flash supported), Cursor/Claude with MCP support.
• Demo: https://github.com/user-attachments/assets/6450b5b7-10e5-4019-82a4-6d726dbfbe1f

Highlighted Details

• Tests live websites and localhost development sites.
• Supports specifying URL, number of agents, and headless mode.
• Can run with multiple agents for thorough testing.
• Leverages Browser-Use technology.

Maintenance & Community

No specific information on contributors, sponsorships, or community channels is provided in the README.

Licensing & Compatibility

• License: MIT
• Compatible with commercial use and closed-source linking due to the permissive MIT license.

Limitations & Caveats

The project requires a specific IDE (Cursor/Claude) with MCP support for full integration. A Google API key is mandatory for operation.