Web UX Evaluator

The MCP Server of operative.sh is an LLM-powered autonomous debugging tool that can execute and debug web applications directly in the code editor.

Developer tools Browser automation #Autonomous debugging #Web testing #IDE integration .Python

rating : 0 points

downloads : 537

update time : 2025-04-28

What is the web-eval-agent MCP Server?

An intelligent debugging assistant that uses AI to automatically test and debug web applications right in your code editor. It works seamlessly with IDEs like Cursor to identify and fix issues in real-time.

How does it work?

The agent autonomously navigates your webapp, captures network traffic and console errors, then provides actionable feedback - all without manual testing.

When should I use it?

Ideal for developers who want to: - Save time on manual testing - Catch bugs before deployment - Get AI-assisted debugging - Streamline web development workflow

Key Features

Autonomous Web NavigationThe AI agent automatically interacts with your webapp 2x faster using operative's optimized backend

Network Traffic MonitoringCaptures all requests and responses for complete visibility

Console Error DetectionAutomatically surfaces all JavaScript errors and warnings

Integrated DebuggingWorks directly within Cursor/Cline IDEs to test code changes in real-time

Pros and Cons

Advantages

Saves hours of manual testing time

Catches issues human testers might miss

Seamless integration with development workflow

Free to get started with basic features

Limitations

Requires initial setup (brew/npm)

Windows installation currently manual

May require IDE restart after installation

Getting Started

Install Prerequisites

Ensure you have Homebrew, npm and jq installed

Run Installer

Download and execute the installation script (you can inspect it first if preferred)

Activate in IDE

Enable the agent in Cursor's Agent Mode (may require IDE refresh/restart)

Usage Scenarios

Automated Form ValidationThe agent automatically tests form submissions and validation rules

API Response VerificationThe agent monitors API calls and verifies response formats

Frequently Asked Questions

Is an API key required?

How do I install on Windows?

What if I get Playwright errors?

Additional Resources

Official Website

Get your API key and learn more

GitHub Repository

Source code and issue tracking

Installation Script

View the installation script before running

🚀 operative.sh web-eval-agent MCP Server

Let the code fix itself. You've got more important things to do.

Demo

🚀 Quick Start

💻 macOS/Linux Installation

Prerequisites: Install brew with /bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)", npm using brew install npm, and jq with brew install jq.
Obtain a free API key from here and then run the installer:

# Feel free to review the installation script content as follows:
# curl -LSf https:// operative.sh/install.sh | less -N
# Download, install, and remove the installation script
curl -LSf https:// operative.sh/install.sh -o install.sh && bash install.sh && rm install.sh

Launch the agent in Cursor Agent Mode using web_eval_agent (refresh the tool or restart Cursor).
If you encounter any issues, refer to the troubleshooting section below.

🪟 Manual Windows Installation (Cline)

# 1. Get your API key from operative.sh
# 2. Install uv (curl -LsSf https://astral.sh/uv/install.sh | sh)
# 3. uvx --from git+https://github.com/Operative-Sh/web-eval-agent.git playwright install
# 4. Start the agent with web_eval_agent in Cline (you may need to restart Cline)

✨ Features

These tools bring a new experience to your code IDE (Cline, Cursor):

🌐 Navigate your web app with BrowserUse (2x faster with the operative backend).
📊 Capture network traffic to keep track of all requests/responses.
🚨 Collect console errors so they can't hide.
🤖 Autonomous debugging: The Cursor agent calls the web QA agent mcp server to test if the code functions end - to - end as expected.

🔥 Accelerate Your Debugging Experience

The MCP server of operative.sh executes and debugs web applications directly in your code editor through an LLM (Large Language Model) - driven agent.

🚨 Important Note

Issues that might occur when the tool calls Playwright for the first time have been fixed on 4/14. Run npm install -g playwright to resolve Playwright - related problems.
If you have any issues, feel free to submit an Issue in this repository!

Made with ❤️ by @ operative.sh