Simple MCP Server to enable a human-in-the-loop workflow in tools like Cline and Cursor.
https://github.com/mrexodia/user-feedback-mcpYou know the frustration: your AI assistant just spent 20 minutes implementing a complex UI feature, claims it's perfect, but you have no idea if it actually works until you manually test it. Then you discover the button doesn't click, the styling is broken, or the workflow makes no sense.
User Feedback MCP solves this by letting your AI assistant pause and ask for your input before moving on. No more blind development cycles.
When you're building desktop applications with Cline or Cursor, your AI assistant is essentially coding blind. It can write the code, but it can't:
This leads to endless back-and-forth cycles where you test, report issues, and wait for fixes.
This MCP server creates a feedback checkpoint in your development workflow. Your AI assistant can now:
✅ Implement the feature
✅ Pause and ask: "I've added the login form. Can you test it and let me know if the validation works correctly?"
✅ Wait for your feedback
✅ Continue with fixes or move to the next task
The assistant gets real-time feedback instead of making assumptions, dramatically reducing the revision cycles.
Before:
After:
Desktop Application Development
"I've implemented the drag-and-drop file upload. Can you try dragging a few files and tell me if the progress indicators work correctly?"
Complex UI Components
"I've built the multi-step wizard form. Can you walk through it and let me know if the navigation between steps feels smooth?"
Integration Testing
"I've connected the payment processing. Can you test a transaction and confirm the success/error states display properly?"
The .user-feedback.json file lets you automate common testing commands:
{
"command": "npm run dev",
"execute_automatically": true
}
When the AI asks for feedback, your development server starts automatically. No manual npm run dev every time.
For complex testing workflows, integrate with task runners:
{
"command": "task test:ui",
"execute_automatically": true
}
Clone the repo and add this to your Cline MCP configuration:
{
"mcpServers": {
"github.com/mrexodia/user-feedback-mcp": {
"command": "uv",
"args": [
"--directory", "path/to/user-feedback-mcp",
"run", "server.py"
],
"autoApprove": ["user_feedback"]
}
}
}
Add this to your custom prompt for best results:
Before completing the task, use the user_feedback MCP tool to ask the user for feedback.
Instead of your AI assistant working in isolation and hoping for the best, it becomes a collaborative partner that knows when to pause and get your input. You catch issues immediately instead of discovering them after the assistant has already moved on to the next feature.
For desktop and UI development where visual confirmation is critical, this turns your AI assistant from a blind coder into an intelligent collaborator that actually understands what "working correctly" means.
The difference is dramatic: fewer revision cycles, faster development, and AI assistance that actually understands your project's real-world requirements.