Interactive Feedback

Interactive Feedback MCP is a high-performance MCP Server designed to streamline AI-assisted development by enabling real-time, bidirectional interaction between your AI tools and human input. It helps you clarify requirements, reduces unnecessary API usage, and accelerates iteration with smart analysis and predefined options while delivering a modern, responsive UI that works across tools like Cursor, Claude Desktop, and Windsurf.

How to use

You use Interactive Feedback MCP by connecting it to an MCP client through a local stdio interface or a remote HTTP endpoint. The server provides intelligent interaction during development, offering real-time clarification prompts, quick-selection options, and project-aware recommendations. Your AI assistant can pause to confirm actions, ensuring you only execute operations after your explicit approval. Use the built-in performance monitoring to keep an eye on startup time, UI responsiveness, and memory usage as you work.

Typical usage patterns include initiating a project analysis to surface improvement suggestions, asking for clarification when requirements are unclear, and performing batch-style or complex actions through the predefined option set. You can switch between themes and languages for the UI, and leverage the three-column layout to review messages, recommendations, and project status side by side.

How to install

Prerequisites: ensure you have Python 3.10 or newer and a supported operating system (Windows 10+, macOS 12+, Ubuntu 20.04+). You should have at least 4 GB of memory and 500 MB of free disk space.

1. Install the uv package manager by running the following command in a terminal.

# macOS/Linux
curl -LsSf https://astral.sh/uv/install.sh | sh

# Windows (PowerShell)
powershell -c "irm https://astral.sh/uv/install.ps1 | iex"

# Alternatively, install via pip
pip install uv

2. Clone the Interactive Feedback MCP repository and navigate into the project directory.

git clone https://github.com/your-username/interactive-feedback-mcp.git
cd interactive-feedback-mcp

3. Verify the installation by starting the MCP server and launching the UI for a quick test.

# Start the MCP server
uv run server.py

# Test the UI with a sample prompt and output file
uv run enhanced_feedback_ui.py --prompt "测试消息" --output-file test.json

Configuration and usage details

Configure MCP clients to connect to Interactive Feedback MCP. The following examples show how to register the MCP server on two clients: Cursor and Claude Desktop.

# Cursor MCP client configuration (mcp.json)
{
  "mcpServers": {
    "interactive_feedback": {
      "command": "uv",
      "args": [
        "--directory", "/absolute/path/to/interactive-feedback-mcp",
        "run",
        "server.py"
      ],
      "timeout": 600,
      "autoApprove": ["interactive_feedback"]
    }
  }
}

For Claude Desktop, edit the configuration so the client can start the MCP server with the same runtime command.

# Claude Desktop configuration (claude_desktop_config.json)
{
  "mcpServers": {
    "interactive_feedback": {
      "command": "uv",
      "args": [
        "--directory", "/absolute/path/to/interactive-feedback-mcp",
        "run",
        "server.py"
      ],
      "timeout": 600
    }
  }
}

AI assistant rules and performance goals

Set up rules to ensure interactive usage, including mandatory confirmation before executing actions and providing predefined options for common operations.

Performance goals to monitor and maintain:

{
  "ui": {
    "theme": "enhanced_glassmorphism",
    "language": "zh_CN",
    "font_family": "PingFang SC",
    "font_size": 14,
    "window_width": 1400,
    "window_height": 1200,
    "panel_ratios": [40, 40, 20]
  },
  "performance": {
    "max_startup_time": 2.0,
    "max_response_time": 100.0,
    "max_memory_usage": 100.0,
    "enable_monitoring": true
  }
}

Logging and troubleshooting

A comprehensive logging system records multiple log levels, supports file rotation, and captures performance metrics and project context. Use the provided tooling to view, search, monitor, and export logs as needed.

Key log locations include main, error, performance, and project context logs.

Common troubleshooting steps include checking for encoding issues, resetting configuration, and verifying MCP connections.

Available tools

interactive_feedback

A fast, interactive feedback tool that analyzes user intent, offers predefined options, and pauses to confirm actions before execution.

manage_logs

Utility for viewing, summarizing, searching, and exporting log files to diagnose issues and track performance.