Restaurant Booking MCP Server
An AI-powered Model Context Protocol (MCP) server for restaurant discovery and booking. This server integrates with Google Maps Places API to find restaurants based on location, cuisine preferences, mood, and event type, then provides intelligent recommendations and booking assistance.
🎯 Key Features
- Smart Restaurant Search: Find restaurants within 20km radius with advanced filtering
- Default Taiwan Location: Automatically searches around Taiwan (24.1501164, 120.6692299) when no coordinates specified
- AI-Powered Recommendations: Get top 3 restaurant suggestions with detailed reasoning
- Google Maps Integration: Real restaurant data including ratings, reviews, and photos
- Event-Specific Matching: Optimized for dating, family gatherings, business meetings, and celebrations
- Mood-Based Filtering: Find restaurants matching romantic, casual, upscale, fun, or quiet atmospheres
- Booking Assistance: Get reservation instructions and mock booking capabilities
Features
- 🔍 Smart Restaurant Search: Find restaurants within 20km radius based on location, cuisine types, mood, and event type
- 📍 Google Maps Integration: Real restaurant data with ratings, reviews, photos, and contact information
- 📅 Booking Assistance: Check availability and get reservation instructions
- 🎯 Event-Specific Matching: Optimized recommendations for dating, family gatherings, business meetings, etc.
- 🎭 Mood-Based Filtering: Find restaurants that match your desired atmosphere (romantic, casual, upscale, etc.)
Prerequisites
- Node.js 18+
- Google Maps API Key with Places API enabled
- TypeScript knowledge for customization
Installation
-
Clone or download this project
git clone <repository-url> cd mcp-restaurant-booking -
Install dependencies
npm install -
Set up environment variables
cp .env.example .envEdit
.envand add your Google Maps API key:GOOGLE_MAPS_API_KEY=your_actual_api_key_here -
Build the project
npm run build
Getting Google Maps API Key
- Go to Google Cloud Console
- Create a new project or select existing one
- Enable the following APIs:
- Places API
- Maps JavaScript API
- Geolocation API
- Places API (New)
- Geocoding API
- Create credentials (API Key)
- Restrict the API key to the enabled APIs for security
Usage
Running the Server
Development mode:
npm run dev
Production mode:
npm start
Running in Docker
To run the MCP Restaurant Booking server in Docker:
# Build the Docker image
docker build -t mcp/booking .
# Run the container on the same network as Redis
docker run --rm -i mcp/booking
Available Tools
The MCP server provides the following tools:
1. search_restaurants
Find restaurants based on location, cuisine, mood, and event type.
Parameters:
latitude(number, optional): Search latitude (default: 24.1501164 - Taiwan)longitude(number, optional): Search longitude (default: 120.6692299 - Taiwan)placeName(string, optional): Place name to search near (e.g., "New York", "Tokyo", "London"). Alternative to providing latitude/longitude coordinates.cuisineTypes(string[]): Array of cuisine preferencesmood(string): Desired atmosphereevent(string): Type of occasionradius(number, optional): Search radius in meters (default: 20000)priceLevel(number, optional): Price preference (1-4)
Example with default Taiwan location:
{
"cuisineTypes": ["Chinese", "Taiwanese"],
"mood": "casual",
"event": "family gathering",
"priceLevel": 2
}
Example with explicit coordinates (Taipei):
{
"latitude": 25.033,
"longitude": 121.5654,
"cuisineTypes": ["Italian", "Mediterranean"],
"mood": "romantic",
"event": "dating",
"radius": 15000,
"priceLevel": 3
}
Example with place name (New York):
{
"placeName": "New York, NY",
"cuisineTypes": ["Italian", "American"],
"mood": "upscale",
"event": "business meeting",
"radius": 10000,
"priceLevel": 3
}
Example with keyword search for specific food types:
{
"keyword": "hotpot",
"mood": "casual",
"event": "family gathering",
"radius": 10000
}
2. get_restaurant_details
Get detailed information about a specific restaurant.
Parameters:
placeId(string): Google Places ID of the restaurant
3. get_booking_instructions
Get instructions on how to make a reservation.
Parameters:
placeId(string): Google Places ID of the restaurant
4. check_availability
Check availability for a reservation (mock implementation).
Parameters:
placeId(string): Google Places IDdateTime(string): Preferred date/time in ISO formatpartySize(number): Number of people
5. make_reservation
Attempt to make a reservation (mock implementation).
Parameters:
placeId(string): Google Places IDpartySize(number): Number of peoplepreferredDateTime(string): ISO format date/timecontactName(string): Name for reservationcontactPhone(string): Phone numbercontactEmail(string, optional): Email addressspecialRequests(string, optional): Special requests
How It Works
1. Restaurant Discovery
- Uses Google Places Nearby Search API to find restaurants within specified radius
- Filters by cuisine types using keyword matching
- Retrieves detailed information for each restaurant
2. AI Recommendation Engine
The recommendation system scores restaurants based on:
- Rating & Reviews (40% weight): Higher ratings and more reviews = better score
- Review Count (20% weight): More reviews indicate reliability
- Cuisine Match (20% weight): How well restaurant cuisine matches preferences
- Event Suitability (10% weight): Appropriateness for the specified event type
- Mood Match (10% weight): Atmosphere alignment with desired mood
3. Event-Specific Scoring
Different events have different criteria:
- Dating: Prefers mid-to-high-end, romantic cuisines, avoids fast food
- Family Gathering: Prefers family-friendly, budget-to-mid-range options
- Business Meeting: Prefers quiet, professional, upscale environments
- Casual Dining: Flexible criteria, budget-friendly options
- Celebration: Prefers high-end, special occasion venues
4. Mood Matching
Analyzes restaurant names, reviews, and characteristics for mood keywords:
- Romantic: intimate, cozy, candlelit, wine
- Casual: relaxed, friendly, laid-back
- Upscale: elegant, sophisticated, fine dining
- Fun: lively, energetic, vibrant
- Quiet: peaceful, serene, calm
Development
Project Structure
src/
├── types/ # TypeScript type definitions
├── services/ # Core business logic
│ ├── googleMapsService.ts # Google Maps API integration
│ ├── restaurantRecommendationService.ts # AI recommendation engine
│ └── bookingService.ts # Booking logic (mock)
└── index.ts # MCP server implementation
Scripts
npm run build: Compile TypeScriptnpm run dev: Run in development mode with hot reloadnpm start: Run compiled versionnpm run lint: Run ESLintnpm run lint:fix: Fix ESLint issues
Customization
Adding New Cuisine Types
Edit the cuisineMap in src/services/googleMapsService.ts:
const cuisineMap: { [key: string]: string } = {
new_cuisine_type: "Display Name",
// ... existing mappings
};
Modifying Recommendation Logic
Update scoring algorithms in src/services/restaurantRecommendationService.ts:
calculateRestaurantScore(): Overall scoring logiccalculateEventSuitability(): Event-specific criteriacalculateMoodMatch(): Mood matching logic
Adding New Event Types
- Update the
eventenum insrc/types/index.ts - Add event criteria in
calculateEventSuitability()method
Limitations
- Booking: Currently uses mock implementation. Real booking requires integration with restaurant-specific systems or third-party services like OpenTable
- API Quotas: Google Places API has usage limits and costs
- Real-time Data: Restaurant hours and availability may not be real-time
- Geographic Coverage: Limited to areas covered by Google Places API
Contributing
- Fork the repository
- Create a feature branch
- Make your changes
- Add tests if applicable
- Submit a pull request
License
MIT License - see LICENSE file for details
Support
For issues and questions:
- Check the Google Maps API documentation
- Verify your API key has proper permissions
- Check API quotas and billing
- Review server logs for error details
Future Enhancements
- Real booking system integration (OpenTable, Resy, etc.)
- User preference learning
- Multi-language support
- Advanced filtering (dietary restrictions, accessibility)
- Integration with calendar systems
- Price comparison features
- Social features (reviews, sharing)
Additional Browser Control
Using Browser MCP
- https://chromewebstore.google.com/detail/browser-mcp-automate-your/bjfgambnhccakkhmkepdoekmckoijdlc
- https://docs.browsermcp.io/setup-server#cursor
Sample
- Prompt: - While searching restaurants, please perform as professional personal assistant to evaluate the condition I provided, do not ask too many questions for me to choose, pick the best suitable selection for me, checking the reservation options and guide how to do the reservation. also list down the Signature Dishes from that restaurant and Approximately pricing per person. When booking info has booking url using external url, use the mcp browse tool to work and find reservation steps.
- can you help me book a restaurant nearby hongkong 太平洋廣場, I want to have a date with my wife within a fine-dining at evening 6pm. cost is not a concern and needs to be romatic