Interactive Feedback

Interactive Feedback MCP is a Node.js-based server that provides a human-in-the-loop workflow for AI development tools. It lets you run system commands, view live output, and feed textual feedback to an AI, all through a responsive web UI. It supports per-project configurations and can auto-execute commands when needed to streamline your AI-assisted workflows.

How to use

You use Interactive Feedback MCP by connecting an MCP client (such as Cursor, Claude Desktop, or Cline/Windsurf) to the local server. Start the MCP server first, then configure your client to point at the local process. When a prompt requires human input, open the web UI to review the project, run commands to verify results, and provide feedback to the AI. The server supports real-time updates via WebSocket and per-project configurations so you can tailor the experience for each project.

How to install

Prerequisites: you need Node.js 18.0.0 or newer and npm 8.0.0 or newer.

# 1. Clone the repository
# Replace the URL with the actual repository you’re using
git clone https://github.com/zivhdinfo/interactive-feedback-mcp-nodejs.git
cd interactive-feedback-mcp-nodejs

# 2. Install dependencies
npm install

# 3. (Optional) Configure Speech-to-Text by creating an environment file
cp .env.example .env
# Edit .env and set your API keys if you plan to use OpenAI Whisper, etc.
# OPENAI_API_KEY=your_openai_api_key_here

# 4. Start the MCP server
node server.js

Configuration and usage notes

This MCP server is designed to work with a variety of MCP clients. The typical setup involves running the server locally and then configuring the client to connect to the local stdio-based server process.

{
  "mcpServers": {
    "interactive_feedback": {
      "type": "stdio",
      "command": "node",
      "args": ["server.js"],
      "cwd": "C:/path/to/interactive-feedback-mcp-nodejs",
      "env": {
        "NODE_ENV": "production"
      }
    }
  }
}

Security and environment

Keep your API keys and secrets in environment variables. Do not embed sensitive values in code or configuration files that are committed to version control. If you enable Speech-to-Text features, ensure your keys are stored securely in the environment and are not exposed in client-side code.

Troubleshooting

Common issues include port conflicts (the server uses a fixed port for local MCP connections), permission errors, and WebSocket connectivity problems. If you run into problems, start in debug mode to gather more information.

# Enable debug logging
DEBUG=* node server.js

Notes on usage with MCP clients

The server exposes a straightforward tool interface. You can initiate the interactive feedback flow from your AI assistant, which will then open the web UI for you to review prompts, execute commands, and submit feedback to guide the assistant.

Appendix: Tools exposed by this MCP server

The server provides an interactive tool called interactive_feedback for coordinating project work and collecting user feedback. This tool accepts a project directory and a summary and returns command logs and user feedback.

Example workflow

// AI assistant calls this tool
{
  "tool": "interactive_feedback",
  "arguments": {
    "project_directory": "/path/to/project",
    "summary": "Need feedback on the new feature implementation"
  }
}

Available tools

interactive_feedback

Tool that coordinates project work by running commands, collecting command logs, and receiving user feedback to guide the AI assistant