Introduction#
MCP, or Model Context Protocol, is an open standard and framework that provides a unified way for large language models (LLMs) to connect to and interact with external data sources and tools.
This documentation show how to use Amadeus Discover MCP server and describes the tooling available.
Amadeus Discover exposes some tools under this protocol to ease customers integration with Amadeus APIs from AI agents and LLMs.
Currently, Amadeus Discover exposes 1 tool:
Get recommended activities
An intelligent tool that automatically selects the most appropriate API endpoint based on your request complexity:
Simple location-based requests (latitude/longitude + basic filters like radius, currency, maxRecommendations) are routed to the optimized top-activities endpoint for faster responses with curated top recommendations
Advanced personalized requests (including interests, travelerType, tripContext, or any preference parameters) are routed to the products search endpoint for detailed matching and personalization
Automatic fallback mechanism ensures you always get results - if the top-activities endpoint returns no results or encounters an error, the tool automatically retries using the products endpoint
The tool supports filtering by geographic coordinates (latitude, longitude), search radius in km, user interests, traveler type, trip context, and travel date.
Additional filters include currency and preference weights (pricePreference, tripContextPreference, travelerTypePreference, interestPreference, ratingPreference, seasonPreference).
Results are tailored to match the provided criteria, ensuring relevance and personalization.
The response’s parameter bookingUrl provides a direct link to book the activity.
The response’s parameter travelerUrl provides a link to Amadeus Discover whitelabel page to allow the activity to be booked by a travel agency agent.
This intelligent tool is designed to help agents recommend activities that align with the user’s preferences and travel plans. It can recommend the top activities to do at destination or fully customized activities while optimizing performance based on the information provided.