Learn how the Booking.com MCP server connects AI applications with real-time inventory data.
Built on the open Model Context Protocol (MCP), the Booking.com MCP server connects your AI applications and agents to Booking.com's live data. Unlike traditional REST APIs that require multiple endpoints and orchestration, the MCP server provides a unified interface designed for Large Language Models (LLMs) and AI assistants.
MCP can have a range of benefits for different audiences:
- For developers - Simplifies AI integrations by removing endpoint orchestration and handling search complexity for you.
- For AI applications/agents - Provides structured, real-time travel data optimised for LLMs.
- For end-users - Enables more capable assistants that can search and compare accommodations and ancillary services through natural language.
✅ Dynamic tool discovery - MCP clients discover available tools and their live schemas automatically. The tools available are scoped to your affiliate ID.
✅ Optimised for AI applications - Data structures and responses are designed for LLM consumption.
✅ Destination handling - The destination parameter accepts location names such as cities, airport IATA codes (e.g., "JFK"), hotel names, landmark names, or district names. The server resolves these location names into the appropriate destination data before searching for matching accommodations.
✅ Stateless and scalable - Every request is independent. There is no session management or state to maintain.
✅ Real-time data - The server provides access to live availability, pricing, and property details from Booking.com's inventory.
The server is designed for any application that uses MCP, including:
AI travel assistants - Build assistants that understand user intent and recommend accommodations based on natural language queries.
Conversational booking agents - Create chat-based interfaces where users can search for properties using conversational language, such as "Find me a pet-friendly hotel in Barcelona under €150."
Voice-activated search interfaces - Develop hands-free travel search for smart speakers, in-car assistants, or mobile voice apps.
Multi-modal travel apps - Combine text, voice, and visual search in travel applications.
LLM-driven search wrappers – Using the MCP server within custom search functions or agent-based pipelines.
These use cases can be combined or extended as part of broader conversational travel experiences.
The accommodations_search tool provides access to Booking.com's accommodation data. Its capabilities include:
Flexible destination search - Accepts location names for cities, regions, landmarks, districts, airport codes, and hotels. Examples include "Manhattan", "Big Ben", "JFK", or "Grand Hotel Vienna".
Date and guest handling - Defines check-in and check-out dates in ISO 8601 format and specifies the number of adults, children, and rooms, including multi-room searches.
Advanced filtering - Filters by price range, property types (hotels, apartments, etc.), facilities (Wi-Fi, parking, pool), and star ratings.
Sorting - Sorts results by price, guest review scores, or relevance.
Content and metadata - Provides real-time pricing, property photos, amenity lists, guest reviews, location data, and multi-language descriptions.
Before getting started, ensure you meet these prerequisites:
Account and partnership - An active partnership agreement with Booking.com is required.
- This provides access to the Booking inventory and credentials for MCP integration.
- Each partner is assigned an Affiliate ID to access the tool.
Technical requirements
- You need an MCP-compatible AI application or agent (see Get started).
- The client application must implement a Model Context Protocol (MCP). If you use the widget UI provided by Booking, it must also support the MCP Apps protocol extension.
- Available frameworks include the MCP SDK for TypeScript/JavaScript and community-maintained clients for other languages.
Knowledge prerequisites - Familiarity with LLMs and their use of tools and APIs is recommended.
- A basic understanding of the Model Context Protocol specification is also helpful.
To integrate with the Booking.com MCP server:
- Work with your Account Manager directly to complete onboarding for your AI use case
- After onboarding you will receive the following credentials:
- Bearer token - An authentication token for your server requests and configuration.
- Affiliate-Id - Your account identifier.
Check the guidelines to get started
Get started
Learn how to configure third-party clients needed to connect with our MCP server.
Get started
Implementation guide
Follow this guide to implement the MCP server together with the accommodations_search tool.
Read now