Seamless interactive prompts directly in VS Code for AI assistants
Interactive MCP
Transform your AI conversations with seamless interactive prompts directly in VS Code. Never lose your flow when AI assistants need user input!

🚀 What This Project Does
When working with AI assistants, sometimes they need to ask you questions or get your input. Instead of interrupting your conversation, this system shows beautiful popups right in VS Code where you can respond seamlessly.
📦 Components
This repository contains two main components:
🎯 VS Code Extension (`interactive-vscode-extension/`)
A VS Code extension that displays interactive popups and connects to the MCP server.
Compatible with:
- VS Code
- Cursor
- Windsurf
- Other VS Code-based editors
🔧 MCP Server (`interactive-mcp-server/`)
A Model Context Protocol server that enables AI assistants to request user input through the extension.
🎬 How It Works
AI Assistant → MCP Server → VS Code Extension → Interactive Popups
- Install the VS Code extension
- Configure your AI assistant to use the MCP server
- Start chatting with your AI assistant
- When input is needed, a popup appears in VS Code
- Your response goes directly back to the AI conversation
✨ Features
- 🎨 Beautiful Interface: Modern popups that match your editor theme
- 🔔 Smart Notifications: Gentle audio chimes and visual cues
- ⚡ Zero Setup: Works immediately after installation
- 🔒 Privacy First: Everything runs locally on your machine
- 🎯 Multiple Input Types:
- Button selection (multiple choice)
- Text input with custom responses
- Yes/No confirmation dialogs
What's New in Version 0.3.3
- Fixed image display in VS Code Extensions Marketplace
- Improved README presentation with better visual examples
- Connection in multiple IDEs is now possible, enabling multi-instance and multi-workspace support
🚀 Quick Start
For Users
📦 **Step 1: Install the Extension**
- Open VS Code, Cursor, Windsurf, or any VS Code-based editor
- Go to Extensions (Ctrl+Shift+X / Cmd+Shift+X)
- Search for "Interactive MCP"
- Click Install
⚙️ **Step 2: Get MCP Configuration**
After installation, you need to get the configuration to add to your AI assistant's MCP setup:
Option A: Welcome Notification (Recommended)
- Look for a notification that says "Interactive MCP installed successfully!"
- Click "Copy MCP JSON" button in the notification
- ⚠️ Note: This notification only appears once after installation
Option B: Command Palette (Always Available)
- Press
Ctrl+Shift+P(Windows/Linux) orCmd+Shift+P(macOS) - Type: "Interactive MCP: Copy MCP JSON Configuration"
- Press Enter - the configuration is now copied to your clipboard
🔧 **Step 3: Configure Your AI Assistant**
The configuration depends on which AI assistant you're using. Here's the general process:
Find your AI assistant's MCP configuration file (varies by assistant)
Add the MCP server configuration:
If this is your first MCP server:
For VS Code, Cursor, and most AI assistants:
{ "mcpServers": { "interactive-mcp": { "url": "http://localhost:8090/mcp" } } }For Windsurf:
{ "mcpServers": { "interactive-mcp": { "serverUrl": "http://localhost:8090/mcp" } } }If you already have other MCP servers:
For Github Copilot, Cursor, and most AI assistants:
{ "mcpServers": { "existing-server": { // your existing server config }, "interactive-mcp": { "url": "http://localhost:8090/mcp" } } }For Windsurf:
{ "mcpServers": { "existing-server": { // your existing server config }, "interactive-mcp": { "serverUrl": "http://localhost:8090/mcp" } } }Save the configuration file
Restart your IDE completely
🌐 MCP Configuration: The extension uses HTTP transport for a simple, update-resistant configuration:
For Github Copilot, Cursor, and most AI assistants:
{
"mcpServers": {
"interactive-mcp": {
"url": "http://localhost:8090/mcp"
}
}
}
For Windsurf:
{
"mcpServers": {
"interactive-mcp": {
"serverUrl": "http://localhost:8090/mcp"
}
}
}
This configuration won't break when you update the extension! See HTTP-TRANSPORT-README.md for details.
🔌 **Step 4: Connect the Extension**
- Check the status bar at the bottom of your editor
- Look for the "Interactive MCP" indicator:
- "🚫 Interactive MCP Tools Off" - Click it to connect
- "✔️✔️ Interactive MCP Tools Ready" - You're re
Configuration
{"mcpServers": {"interactive-mcp": {"url": "http://localhost:8090/mcp"}}}