Tripadvisor MCP Server

The Tripadvisor MCP server is a community-built (not official) Model Context Protocol server that wraps the Tripadvisor Content API. It exposes a small set of read-only tools that let an AI assistant search for travel locations, look up detailed information, and retrieve reviews and photos for hotels, restaurants, and attractions. Authentication uses a Tripadvisor Content API key, which must be obtained from the Tripadvisor Developer Portal.

The server is built in Python and managed with the uv package manager. It ships with both a local uv run workflow and a Docker setup, and is configured for Claude Desktop out of the box. The list of exposed tools is configurable so operators can limit which capabilities are available to clients.

Note: this is a community project (pab1it0/tripadvisor-mcp) and is not maintained by Tripadvisor or Anthropic. It is read-only: there are no booking, write, or transactional capabilities, and usage is bounded by the rate limits and terms of the Tripadvisor Content API.

Tools

Prerequisites

• A Tripadvisor Content API key from the Tripadvisor Developer Portal
• Python with the uv package manager installed

Install uv

curl -LsSf https://astral.sh/uv/install.sh | sh

Clone and set up the project

git clone https://github.com/pab1it0/tripadvisor-mcp.git
cd tripadvisor-mcp
uv venv && source .venv/bin/activate && uv pip install -e .

Environment variables

TRIPADVISOR_API_KEY=your_api_key_here

Claude Desktop config

{
  "mcpServers": {
    "tripadvisor": {
      "command": "uv",
      "args": ["--directory", "<path-to-repo>", "run", "src/tripadvisor_mcp/main.py"],
      "env": {
        "TRIPADVISOR_API_KEY": "your_api_key_here"
      }
    }
  }
}

Docker (optional)

docker build -t tripadvisor-mcp-server .

Run with docker run passing TRIPADVISOR_API_KEY as an environment variable, or use the included Docker Compose setup with a .env file.

• Build a travel research agent that finds top-rated restaurants near a user's hotel using search_nearby_locations and get_location_details.
• Summarize sentiment from recent reviews of a target hotel or attraction by calling get_location_reviews and aggregating themes.
• Generate richly illustrated itineraries by pairing search_locations with get_location_photos to surface visual context for each stop.
• Power a "should I visit?" assistant that compares multiple attractions in a city by querying details and reviews for each candidate.
• Enrich a CRM or trip-planning tool with Tripadvisor metadata (category, ratings, address) for venues a user has saved.

• "Find the top-rated sushi restaurants within 1 km of latitude 35.6595, longitude 139.7005 in Tokyo."
• "Get details and recent reviews for the Louvre Museum from Tripadvisor and summarize what visitors complain about."
• "Search Tripadvisor for boutique hotels in Lisbon and show me photos of the top three."
• "Compare reviews for the Park Hyatt and the Mandarin Oriental in New York and tell me which has better service ratings."
• "Pull photos for the Colosseum in Rome and pick five suitable for a travel brochure."

• Covers the core read endpoints of the Tripadvisor Content API (search, details, reviews, photos) in a single MCP server.
• Configurable tool list so operators can restrict which capabilities are exposed to clients.
• Both uv and Docker workflows are documented, including a Claude Desktop config example.
• MIT licensed.

• Community-maintained (pab1it0); not an official Tripadvisor or Anthropic project, so there are no SLAs.
• Read-only: no booking, reservations, or write operations are supported.
• Requires a Tripadvisor Content API key with its own approval process and rate limits.

• hhw67865/tripadvisor-mcp-server: another community Tripadvisor MCP focused on vacation planning.
• GongRzhe/TRAVEL-PLANNER-MCP-Server: general travel planning MCP server.
• skarlekar/mcp_travelassistant: multi-server suite for flights, hotels, events, and itineraries.