vibetest-use 0.1.2

Browser-agent QA swarm with MCP interface for testing AI-generated websites

🚀 Vibetest Use

🤖 AI-Powered Multi-Agent Website Testing Platform

Autonomous browser agents that intelligently test your websites for bugs, accessibility issues, and user experience problems

📦 Installation • 🚀 Quick Start • 📖 Usage • 🎯 Examples • 🔧 Configuration

---

✨ What is Vibetest Use?

Vibetest Use is a cutting-edge Model Control Protocol (MCP) server that orchestrates multiple intelligent browser agents to comprehensively test your websites. Powered by Google's Gemini AI and advanced browser automation, it provides enterprise-grade testing capabilities through a simple, intuitive interface.

🎯 Key Features

• 🤖 Multi-Agent Testing: Deploy multiple AI agents simultaneously for comprehensive coverage
• 🧠 AI-Powered Analysis: Gemini 2.0 Flash provides intelligent bug detection and reporting
• 🌐 Universal Compatibility: Test any website - production, staging, or localhost
• ♿ Accessibility Testing: Automated WCAG compliance checking
• 🔗 Link Validation: Comprehensive broken link detection
• 📱 Responsive Testing: Cross-device compatibility validation
• 🎨 UI/UX Analysis: Visual regression and user experience assessment
• 📊 Detailed Reporting: Comprehensive test results with actionable insights

https://github.com/user-attachments/assets/9558d051-78bc-45fd-8694-9ac80eaf9494

---

📦 Installation

Recommended: Using uvx (fastest)

uvx is the fastest way to run Python tools without installation conflicts:

# Run vibetest-use directly (no installation needed)
uvx vibetest-use

# Or install globally for repeated use
uvx --global vibetest-use

Why uvx?

• ⚡ Instant execution - No virtual environment setup
• 🔒 Isolated dependencies - No conflicts with your system
• 🚀 Always latest version - Automatically uses the newest release

Alternative: Using pip

pip install vibetest-use

Prerequisites

• Python 3.11+ - Modern Python version required
• Google API Key - For Gemini AI integration (Get yours here)
• MCP-Compatible IDE - Claude Desktop, Cursor, or any MCP-enabled environment

---

🚀 Quick Start

Option 1: Claude Desktop (Recommended)

# Add MCP server via Claude CLI
claude mcp add vibetest-use vibetest-use -e GOOGLE_API_KEY="your_api_key"

# Verify connection
claude
> /mcp
  ⎿  MCP Server Status
     • vibetest-use: ✅ connected

Option 2: Cursor IDE

1. Open Cursor Settings → MCP → Add Server
2. Configure the server:

{
  "mcpServers": {
    "vibetest-use": {
      "command": "vibetest-use",
      "env": {
        "GOOGLE_API_KEY": "your_api_key"
      }
    }
  }
}

3. Restart Cursor and start testing!

Option 3: Manual Installation

# Clone and install from source
git clone https://github.com/im47cn/vibetest-use.git
cd vibetest-use
pip install -e .

# Run directly
export GOOGLE_API_KEY="your_api_key"
vibetest-use

---

📖 Usage

Basic Commands

Once configured, you can start testing with simple natural language commands:

🌐 Test any website
> Test my website with vibetest-use: https://example.com

🏠 Test localhost development
> Run vibetest-use on localhost:3000

🚀 Advanced testing with multiple agents
> Run a headless vibetest-use on localhost:8080 with 10 agents

🔍 Focused accessibility testing
> Test accessibility issues with vibetest-use on my staging site: https://staging.myapp.com

Advanced Parameters

Parameter	Description	Example
URL	Any website URL	https://example.com, localhost:3000
Agent Count	Number of testing agents (1-20)	5 agents, 10 agents
Mode	Headless or visual testing	headless, non-headless
Focus	Specific testing areas	accessibility, links, ui

---

🎯 Examples

E-commerce Site Testing

> Test my e-commerce site with 8 agents focusing on checkout flow: https://mystore.com

Mobile App Testing

> Run responsive testing on my mobile app: https://app.example.com

Accessibility Audit

> Perform comprehensive accessibility testing on localhost:3000 with detailed reporting

Performance Testing

> Test loading performance and user experience on https://mysite.com with 5 agents

---

🔧 Configuration

Environment Variables

# Required
export GOOGLE_API_KEY="your_gemini_api_key"

# Optional
export BROWSER_USE_LOGGING_LEVEL="INFO"
export ANONYMIZED_TELEMETRY="false"

Advanced MCP Configuration

For enterprise deployments, you can customize the MCP server configuration:

{
  "mcpServers": {
    "vibetest-use": {
      "command": "vibetest-use",
      "env": {
        "GOOGLE_API_KEY": "your_api_key",
        "MAX_AGENTS": "10",
        "DEFAULT_TIMEOUT": "30",
        "HEADLESS_MODE": "true"
      }
    }
  }
}

---

🎬 Full Demo

https://github.com/user-attachments/assets/6450b5b7-10e5-4019-82a4-6d726dbfbe1f

---

🤝 Contributing

We welcome contributions! Please see our Contributing Guide for details.

Development Setup

git clone https://github.com/im47cn/vibetest-use.git
cd vibetest-use
pip install -e ".[dev]"
pytest

---

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

---

🙏 Acknowledgments

• Powered by Browser Use - Advanced browser automation
• Built with Google Gemini AI - Intelligent analysis and reporting
• Supports Model Control Protocol - Seamless AI integration

---

Made with ❤️ for developers who care about quality

⭐ Star us on GitHub • 🐛 Report Issues • 💬 Join Discussions