Interactive Feedback

Interactive Feedback MCP is a powerful server that enables interactive user feedback and command execution within your development workflow. It integrates with MCP clients to guide AI-assisted tasks, run commands in your project, and handle image inputs and quick replies to keep feedback loop smooth.

How to use

You connect to the Interactive Feedback MCP server through an MCP client like uvx or uv. The server provides an interactive feedback interface that can accept text and image input, execute commands in your project directory, and optionally submit feedback automatically after a countdown.

How to install

Prerequisites you need: a running MCP client tool (uvx or uv) and either Python or a Node-based environment depending on how you want to run the server.

Direct usage (recommended) Start the server without installing anything locally by using the MCP client to run the tool directly:

uvx interactive-feedback-mcp

uvx --from git+https://github.com/duolabmeng6/interactive-feedback-mcp.git interactive-feedback-mcp

Local development options if you want to run the server from a local clone or a local script include the following approaches.

# 1) Use uv to run when you have the project locally
uv run interactive-feedback-mcp

# 2) Run the Python server script directly
uv run python interactive_feedback_mcp/server.py

Additional configuration and usage notes

You can configure how the MCP server is invoked from your MCP client using the following example configurations. These entries show separate start methods you might use depending on whether you prefer uvx, uv, or Python execution.

{
  "mcpServers": {
    "interactive_feedback_mcp": {
      "command": "uvx",
      "args": ["--from", "git+https://github.com/duolabmeng6/interactive-feedback-mcp.git", "interactive-feedback-mcp"],
      "timeout": 600,
      "autoApprove": ["interactive_feedback"]
    }
  }
}

{
  "mcpServers": {
    "interactive_feedback_mcp": {
      "command": "uv",
      "args": ["run", "--directory", "/path/to/interactive-feedback-mcp", "interactive-feedback-mcp"],
      "timeout": 600,
      "autoApprove": ["interactive_feedback"]
    }
  }
}

{
  "mcpServers": {
    "interactive_feedback_mcp": {
      "command": "python",
      "args": ["/path/to/interactive-feedback-mcp/interactive_feedback_mcp/server.py"],
      "timeout": 600,
      "autoApprove": ["interactive_feedback"]
    }
  }
}

Troubleshooting and tips

If you don’t see a graphical interface after starting the server, that is expected behavior until the MCP client invokes interactive feedback. Ensure your MCP client is configured to call the interactive-feedback-mcp tool during the workflow to trigger the UI.

For best results, keep feedback flowing at each key step, leverage image inputs when helpful, and use terminal commands to verify and implement user requests.

许可和支持

This project is released under the MIT license. Use it in accordance with the license terms.

Available tools

interactive_feedback

Provides an interactive feedback interface within MCP for text and image input, including command execution and auto-submission features.

Context7

Queries the latest library documents and examples to assist with running and configuring MCP servers.

MCP service

Represents the overarching MCP service enabling interactive feedback, image handling, and command execution workflow.