home / mcp / ai interaction tool - mcp server

AI Interaction Tool - MCP Server

Provides a local MCP server that enables interactive AI sessions with file and image attachments and tag-based outputs.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "khaihuynhvn-mcp-server_ai-interaction": {
      "command": "python",
      "args": [
        "E:/MCP-servers-github/AI-interaction/mcp_server.py"
      ],
      "env": {
        "PYTHONUNBUFFERED": "1"
      }
    }
  }
}

You run an MCP (Model Context Protocol) server that hosts the AI Interaction Tool. This server lets an MCP client start a UI-based AI interface, attach files and images, and receive structured, tag-based outputs that integrate smoothly with your AI agent rules. It runs locally, processes inputs securely, and supports multi-image attachments with persistent state if you enable it.

How to use

Launch the MCP server from your development environment, then connect to it with an MCP client such as Claude Desktop or a compatible UI client. You will be able to open an interactive popup for user input, attach files or folders from your workspace, add and manage multiple images, and receive a structured, tag-based AI response that includes any attached files in base64 form for seamless processing.

How to install

# Prerequisites
- Python 3.8+ should be installed on your system
- A suitable MCP client (Claude Desktop, Cursor IDE, or equivalent) ready to connect

# Step 1: Clone the project repository
git clone https://github.com/your-username/AI-interaction.git
cd AI-interaction

# Step 2: Set up Python environment and install dependencies
# Using pip (default)
pip install -r requirements.txt

# Optional for performance: install uvx (as described in the setup guide)
pip install uv
uv pip install -r requirements.txt

# Step 3: Run the MCP server (stdio/local runtime)
python mcp_server.py
```

Note: The MCP server is configured to run as a local stdio server. You connect to it from your MCP client using the provided configuration snippet. If you run the UI directly for testing, you can start the UI with:
```
python E:/MCP-servers-github/AI-interaction/mcp_server.py
`

Configuration and usage notes

Configure your MCP client with a local stdio server that points to the mcp_server.py file. The following configuration was provided for Claude Desktop and should be adapted to your environment, using the absolute path to the mcp_server.py on your system.

{
  "mcpServers": {
    "AI_interaction": {
      "command": "python",
      "args": ["E:/MCP-servers-github/AI-interaction/mcp_server.py"],
      "stdio": true,
      "enabled": true
    }
  }
}
```

Replace the path with your actual absolute path to mcp_server.py.

Additional sections

Security and privacy are emphasized: all data processing happens locally, images and files are stored under controlled directories, and there is no external data transmission. Manage image attachments securely with the option to save or auto-cleanup based on your preferences.

If you need to test or debug, you can run the server directly to observe logs and behavior, and verify your client can connect and display the interactive UI as expected.

Troubleshooting and tips

Common issues include connection failures between the MCP client and the local server, missing dependencies, or UI rendering problems. Ensure Python is installed and on your PATH, verify the absolute path to mcp_server.py in your client configuration, install dependencies from requirements.txt, and confirm PyQt5 is available for the UI.

Available tools

ai_interaction

Main interactive MCP tool that opens a UI popup for content input with file/image attachments and outputs a structured, tag-based result ready for AI agent processing.