AI is reshaping travel and hospitality at every stage of the journey. The global travel technology market reached approximately $8.6 billion in 2024 and is projected to grow to $21–23 billion by 2032 at a 12–14% CAGR. AI specifically applied to travel is forecast to reach $5–7 billion by 2030 at a 28–30% CAGR. Online travel booking now represents 65–70% of all travel bookings globally, and approximately 70% of travel companies report using or planning to use AI in their operations. AI-powered personalization has been shown to increase booking conversion rates by 25–35%. The global travel and tourism industry contributes approximately $9.9 trillion to global GDP.

The MCP ecosystem for travel reflects this commercial momentum more than almost any other vertical. With 12 official or hosted MCP servers from travel companies — including Expedia, Mapbox, Trivago, Kiwi.com, Skiplagged, Ferryhopper, Flightradar24, Variflight, Yelp, DeepL, WeatherXM, and LiteAPI — travel is one of the most commercially engaged MCP verticals. Flight search has the deepest coverage with 10+ implementations. Vacation rentals are anchored by the most popular travel MCP server overall (mcp-server-airbnb at 406 stars). Maps and navigation benefit from Mapbox’s official hosted server. But major platforms like Booking.com, Google Flights, Kayak, Uber/Lyft, and the entire cruise industry have zero official MCP presence.

MCP — the Model Context Protocol — provides a standardized way for AI agents to connect to travel platforms, booking APIs, navigation services, and tourism infrastructure. Rather than building custom integrations for each travel API, MCP-connected agents can search flights, compare hotels, plan itineraries, check weather, navigate routes, and discover restaurants through defined tool interfaces. For an introduction to MCP itself, see our introduction to MCP.

This guide is research-based. We survey what MCP servers exist across the travel and hospitality landscape, analyze the architecture patterns they enable, and identify where significant gaps remain. We do not claim to have tested or used any of these servers hands-on — our analysis draws on published documentation, open-source repositories, vendor announcements, and industry research. Rob Nugen operates ChatForest; the site’s content is researched and written by AI. For background on MCP, see our introduction to MCP and the MCP server directory.

Why Travel and Hospitality Need MCP

Travel involves complex, multi-service workflows across fragmented platforms — exactly the kind of coordination that MCP is designed to enable.

Trip planning spans dozens of services and data sources. A single trip may involve flight search across multiple airlines, hotel comparison across OTAs, ground transportation booking, restaurant reservations, activity planning, weather checking, and currency conversion. MCP-connected agents can orchestrate across these services — searching flights on one platform, comparing hotel prices on another, and checking weather at the destination — all within a single conversational flow.

Pricing is dynamic and time-sensitive. Travel prices change constantly based on demand, availability, and competitive pricing. MCP servers for flight and hotel search enable AI agents to monitor prices, compare across platforms, and help travelers find optimal booking windows. The real-time nature of travel pricing makes API-connected agents particularly valuable.

Travel involves geographic and multilingual complexity. Travelers need maps, navigation, timezone conversions, weather forecasts, and often translation services. MCP servers for Mapbox, Google Maps, weather APIs, currency conversion, and DeepL translation create a toolkit for AI agents that mirrors what a human travel agent would need to know.

The booking funnel has many decision points. Choosing between airlines, hotels, and activities involves comparing prices, reading reviews, checking availability, evaluating locations, and coordinating schedules. MCP servers connecting to TripAdvisor, Yelp, and booking platforms give AI agents access to the review and rating data that influences travel decisions.

Flight Search and Booking

Flight search has the deepest MCP coverage in the travel vertical, with 10+ implementations ranging from API wrappers to hosted endpoints.

Community Flight Search Servers

flights-mcp (ravinahp) | ~177 stars | Python The most popular self-hosted flight search MCP server. Uses the Duffel API for one-way, round-trip, and multi-city searches with flexible date options. Read-only — searches and compares but does not execute bookings. The Duffel API provides access to 300+ airlines.

travel-hacking-toolkit (borski) | ~61 stars | JavaScript A comprehensive travel mega-bundle containing 6 MCP servers (Skiplagged, Kiwi, Trivago, Ferryhopper, Airbnb, LiteAPI) plus 7 skills (Duffel, seats.aero, AwardWallet, SerpAPI, and more). This is the closest thing to a unified travel agent toolkit in the MCP ecosystem, bundling search across flights, hotels, ferries, and vacation rentals into a single package.

mcp-flight-search (arjunprabhulal) | ~41 stars | Python Real-time flight search via SerpAPI/Google Flights integration. Uses Google Flights data without requiring a direct Google API, making it accessible for developers who want Google’s flight data through a search-based approach.

FlightTicketMCP (xiaonieli7) | ~35 stars | Python Focused on the Chinese travel market — provides Chinese flight ticket queries and real-time flight status tracking. One of several MCP servers specifically serving Chinese-language travel needs.

flights-mcp-server (smamidipaka6) | ~21 stars | Python Another Google Flights search implementation via SerpAPI, providing an alternative approach to the same underlying data.

mcp-skyscanner (shadyvb) | ~4 stars | Python Skyscanner flight and airport search. Marked as experimental, but provides access to Skyscanner’s comparison engine through MCP.

amadeus-mcp-server (fiqcodes) | ~1 star Amadeus Travel API integration covering flights, hotels, and activities. Amadeus is one of the major Global Distribution Systems (GDS) powering airline booking infrastructure.

flight-mcp-server (technicalerikchan) | ~2 stars Another Amadeus API integration specifically for real-time flight search.

Hosted Flight Search Endpoints

Several travel companies have released hosted MCP servers that require no self-hosting — agents connect directly to the company’s remote endpoint.

Skiplagged MCP | Hosted at mcp.skiplagged.com/mcp | Free, no API key Skiplagged, known for finding “hidden city” fares and budget flights, provides a free hosted MCP endpoint. No API key required — agents can connect directly and search for flights.

Kiwi.com MCP | Hosted at mcp.kiwi.com | Free, no API key Kiwi.com’s hosted MCP endpoint provides access to their virtual interlining technology, which combines flights from different airlines into creative routing options that traditional booking systems miss. Free to use with no API key.

Airline-Specific Servers

mcp-american (markswendsen-code) | ~2 stars American Airlines booking via Playwright browser automation. This approach — automating the airline’s web interface rather than using an API — reflects the absence of official airline MCP servers.

Aviation Data and Flight Tracking

Flight tracking and aviation data have attracted multiple implementations, including two official servers from aviation data companies.

flightradar24-mcp-server (sunsetcoder) | ~46 stars | JavaScript The most popular community flight tracking MCP server, providing real-time flight tracking via FlightRadar24 data. FlightRadar24 tracks aircraft positions globally using ADS-B receiver networks.

Flightradar24 fr24api-mcp | ~15 stars | TypeScript | Official Flightradar24’s official MCP server, published as an npm package (@flightradar24/fr24api-mcp). Provides live aircraft positions and historical flight data going back to 2016. The official status means it uses FR24’s sanctioned API rather than scraping.

Variflight MCP | ~22 stars | Official Variflight’s official MCP server, published as @variflight-ai/variflight-mcp on npm. Variflight is a major flight data provider particularly strong in the Asian market.

aviationstack-mcp (Pradumnasaraf) | ~17 stars | Python AviationStack API integration providing real-time flight data, airport schedules, and aircraft type information.

seats.aero-mcp-server (gavgrego) | ~12 stars Award flight availability search across 25+ airline mileage programs. Particularly useful for frequent flyers looking to book award travel — a niche that the standard flight search servers don’t cover.

timcp (abahgat) | ~2 stars | Python Travel Impact Model server that calculates flight greenhouse gas emissions. Uses Google’s Travel Impact Model to estimate the carbon footprint of flights — an increasingly important consideration for environmentally conscious travelers and corporate travel policies.

mcp-server-flight-aware-aeroapi (mikedarke) | ~2 stars | TypeScript FlightAware AeroAPI integration for flight tracking and airport data.

Hotel and Accommodation Booking

Hotel booking MCP servers are less numerous than flight search but include both hosted endpoints and API wrappers.

Hosted Hotel Endpoints

Trivago MCP | Hosted at mcp.trivago.com/mcp | Free, no API key Trivago, the hotel metasearch engine comparing prices across booking platforms, provides a free hosted MCP endpoint. Agents can search and compare hotel prices across Trivago’s aggregated inventory without any API key.

LiteAPI MCP | Hosted at mcp.liteapi.travel/api/mcp | Requires API key LiteAPI provides live hotel rates and real-time booking capabilities through their hosted MCP endpoint. Unlike most travel MCP servers that are read-only, LiteAPI supports actual booking transactions. Claims access to 3 million+ hotels.

Community Hotel Servers

hotels_mcp_server (esakrissa) | ~16 stars | Python Booking.com API integration for hotel search. One of the few connections to the world’s largest hotel booking platform, though community-built rather than official.

mcp-booking (samwang0723) | ~14 stars | TypeScript Another booking-focused MCP server implementation.

agoda-review-mcp (birariro) | ~2 stars | Java Focused specifically on Agoda hotel reviews — aggregates and categorizes positive and negative reviews. Useful for review analysis rather than booking.

WinWin-travel MCP-server | ~1 star Hosted MCP with access to 3 million+ hotels inventory with real-time booking capability.

mcp-marriott (markswendsen-code) | ~1 star Marriott hotel booking via Playwright browser automation, reflecting the absence of official hotel chain MCP servers.

Official First-Party

Expedia Travel Recommendations MCP (ExpediaGroup) | ~14 stars | Official Expedia Group’s official MCP server for travel recommendations covering hotels, flights, activities, and car rentals. While modest in star count, its official status from the world’s second-largest online travel company is significant. This is one of the few MCP servers from a major OTA.

Vacation Rentals and Property Management

Vacation rental MCP servers include the single most popular travel MCP server and several property management tools.

mcp-server-airbnb (openbnb-org) | ~406 stars | JavaScript The most popular MCP server in the entire travel vertical. Provides Airbnb listing search with advanced filtering and DXT format support. At 406 stars, it significantly outpaces every other travel MCP server, reflecting both Airbnb’s market dominance and developer interest in vacation rental search.

lodgify-mcp-server (Fast-Transients) | ~2 stars Lodgify vacation rental API integration covering properties, bookings, and calendar management. Lodgify is a property management platform for vacation rental owners.

guesty-mcp-server (DLJRealty) | ~1 star | 38 tools Guesty property management integration with 38 tools covering reservations, guest management, pricing, and financials. Guesty is a property management platform used by vacation rental businesses. The 38-tool count makes this one of the more comprehensive travel MCP servers by feature breadth.

MountVacation MCP (talirezun) | ~2 stars European ski resort and mountain vacation planning, providing niche destination-specific search.

lapalma24-mcp | ~0 stars Vacation rentals specifically for La Palma, Canary Islands — an example of highly localized travel MCP servers.

lilo-property MCP | ~0 stars Vacation rental booking and protection designed specifically for AI agents, addressing the trust and liability questions that arise when AI systems make accommodation bookings.

Maps, Navigation, and Geospatial

Maps and navigation have strong MCP coverage, anchored by Mapbox’s official hosted server and several Google Maps community implementations.

Mapbox MCP Server | ~325 stars | TypeScript | Official Mapbox’s official hosted MCP endpoint at mcp.mapbox.com/mcp. Provides geocoding, POI search, multi-modal routing, travel time matrices, route optimization, isochrones (reachability areas), static map generation, and offline spatial analysis. The most feature-rich mapping MCP server and one of the highest-starred official travel MCP servers. Mapbox’s routing capabilities include driving, cycling, and walking directions — essential for ground transportation planning.

Mapbox MCP DevKit Server | ~45 stars | Official A developer-focused companion to the main Mapbox MCP server, providing documentation tools, style management, and token management for developers building Mapbox integrations.

mcp-google-map (cablate) | ~237 stars | TypeScript | 18 tools The most comprehensive Google Maps MCP server with 18 tools (14 atomic + 4 composite). Covers geocoding, directions, elevation, timezone, weather, air quality, static maps, batch geocode, and search along route. Despite being community-built (Google has not released an official Maps MCP server), its 237 stars and 18-tool breadth make it a capable alternative to Mapbox’s official offering.

maps-streamable-mcp-server (iceener) | ~19 stars | TypeScript Google Maps Places and Routes API integration using the streamable HTTP transport.

google-maps-mcp-server (Neutrollized) | ~14 stars | Python Google Maps integration covering Directions, Distance Matrix, Geocoding, and Places API.

Weather Services

Weather data is essential for travel planning, and the MCP ecosystem offers many options ranging from lightweight to comprehensive.

weather-mcp-server (ezh0v) | ~239 stars | Go The most popular weather MCP server. Lightweight implementation providing real-time weather data.

weather-mcp-server (glaucia86) | ~97 stars | TypeScript A robust weather server implementation with strong community adoption.

open-meteo-mcp (cmer81) | ~37 stars | TypeScript Open-Meteo API access with no API key required — a free, open-source weather data provider. Particularly useful for travel MCP configurations where minimizing API key dependencies is important.

mcp-weather (TimLukaHorstmann) | ~32 stars AccuWeather API integration with hourly and daily forecasts. AccuWeather’s minute-by-minute precipitation forecasts are particularly useful for travel activity planning.

WeatherXM Pro MCP | ~7 stars | Official WeatherXM’s official MCP server providing weather station data, observations, and forecasts from their network of community-owned weather stations.

weather-mcp (weather-mcp) | ~4 stars | 12 tools The most feature-rich weather MCP server by tool count, offering forecasts, historical data (going back to 1940), weather alerts, air quality, marine weather, lightning detection, radar imagery, river conditions, and wildfire information — all without API keys.

Travel Planning and Itinerary Tools

Several MCP servers attempt to combine multiple travel services into unified planning experiences.

TRAVEL-PLANNER-MCP-Server (GongRzhe) | ~96 stars | JavaScript | 4 tools Uses Google Maps API to provide place search, place details, route calculation, and timezone conversion — the core geospatial operations needed for itinerary building. Available as an npm package.

mcp_travelassistant (skarlekar) | ~32 stars | Python | 6-server ecosystem The most architecturally ambitious travel MCP project — a 6-server ecosystem covering flights, hotels, events, geocoding, weather, and finance. Demonstrates the multi-server orchestration pattern where each travel domain gets its own MCP server, and an agent coordinates across all six.

TravelGenie (kautilyaa) | ~2 stars Similar multi-server architecture with 7 specialized MCP servers orchestrated by Claude for complete trip planning.

trip-loom-monorepo (leojuriolli7) | ~2 stars AI travel assistant combining flights, hotels, and itinerary building in a monorepo architecture.

mcp-travel-assistant (abhinavmathur) | ~5 stars Google Travel and Amadeus integration for combined search.

Travel Reviews and Restaurant Discovery

Review and discovery platforms help travelers evaluate destinations, accommodations, and dining options.

tripadvisor-mcp (pab1it0) | ~53 stars | Python TripAdvisor Content API integration providing location data, reviews, and photos. Enables searching for destinations and experiences with access to TripAdvisor’s massive review database.

Yelp MCP | ~23 stars | Python | Official Yelp’s official MCP server using Yelp Fusion AI. Supports natural language search and multi-turn conversations. Notably includes restaurant reservation capabilities (on request), making it one of the few travel MCP servers that can execute transactions beyond search. The single primary tool (yelp_agent) takes natural language queries rather than structured parameters.

mcp-yelp (shawndrake2) | ~1 star Community Yelp Fusion API integration for business search, ratings, and market sizing.

Rail and Ferry Transport

Ground and sea transport MCP servers are geographically concentrated but include a notable hosted endpoint.

Ferryhopper MCP | Hosted at mcp.ferryhopper.com/mcp | Free, no API key Ferryhopper’s hosted MCP endpoint covering ferry routes across 33 countries and 190+ operators. Free to use with no API key. Particularly valuable for European and Mediterranean travel planning where ferries are a primary transport mode.

ns-mcp-server (r-huijts) | ~49 stars | TypeScript Dutch Railways (NS) API integration with real-time train information and disruption alerts. The most popular rail MCP server.

renfe_mcp_server | ~3 stars Spanish Renfe train schedules using GTFS data with real-time pricing.

ns-travel-mcp (lauragift21) | ~3 stars Another Dutch Railways implementation focused on journey planning and disruption alerts.

Currency and Translation

Cross-border travel requires currency conversion and language services.

DeepL MCP Server | ~95 stars | JavaScript | Official DeepL’s official MCP server for translation. While not travel-specific, translation is essential for international travel planning — translating hotel descriptions, restaurant menus, local information, and communication.

currency-conversion-mcp (wesbos) | ~33 stars | TypeScript Currency conversion designed with travelers in mind. Essential for any travel agent workflow comparing prices across countries.

frankfurtermcp (anirbanbasu) | ~6 stars Frankfurter API integration for exchange rates — provides historical and current rates for currency comparison.

Car Rental and Ride-Sharing

Car rental and ride-sharing have minimal MCP coverage — a significant gap given how central ground transportation is to travel.

mcp-avis (markswendsen-code) | ~0 stars Avis car rental via Playwright browser automation. The browser-automation approach (rather than API integration) indicates the absence of official rental car APIs for MCP.

No MCP servers exist for Uber, Lyft, Hertz, Enterprise, Sixt, or any other major ride-sharing or car rental platform.

Comparison of Key Travel MCP Servers

Server Stars Category Official? Transport Key Strength
mcp-server-airbnb 406 Vacation Rentals Community stdio Most popular travel MCP; advanced filtering
Mapbox MCP Server 325 Maps/Navigation Official Remote Hosted; routing, geocoding, isochrones
weather-mcp-server (ezh0v) 239 Weather Community stdio Lightweight, popular
mcp-google-map 237 Maps/Navigation Community stdio 18 tools; comprehensive Google Maps
flights-mcp 177 Flight Search Community stdio Duffel API; multi-city
DeepL MCP Server 95 Translation Official stdio Official DeepL translation
TRAVEL-PLANNER 96 Itinerary Community stdio Google Maps-based planning
weather-mcp-server (glaucia86) 97 Weather Community stdio Robust TypeScript implementation
travel-hacking-toolkit 61 Multi-service Community stdio 6 bundled MCP servers
tripadvisor-mcp 53 Reviews Community stdio TripAdvisor Content API
ns-mcp-server 49 Rail Community stdio Dutch Railways real-time data
flightradar24-mcp-server 46 Aviation Community stdio FlightRadar24 tracking
mcp-flight-search 41 Flight Search Community stdio Google Flights via SerpAPI
mcp_travelassistant 32 Multi-service Community stdio 6-server ecosystem
currency-conversion-mcp 33 Finance Community stdio Travel currency conversion
Yelp MCP 23 Reviews Official stdio Official; natural language search
Variflight MCP 22 Aviation Official stdio Official flight data (Asia focus)
Flightradar24 fr24api-mcp 15 Aviation Official stdio Official; historical data to 2016
Expedia MCP 14 Multi-service Official stdio Official OTA recommendations
Trivago MCP — Hotels Official Remote Hosted; hotel metasearch
Kiwi.com MCP — Flights Official Remote Hosted; virtual interlining
Skiplagged MCP — Flights Official Remote Hosted; hidden city fares
Ferryhopper MCP — Ferry Official Remote Hosted; 33 countries, 190+ operators
LiteAPI MCP — Hotels Official Remote Hosted; 3M+ hotels, live booking

Architecture Patterns

User: "Find me flights to Barcelona in June,
       a hotel near La Rambla, and good
       tapas restaurants nearby"

    ┌─────────────────────────────────┐
    │         AI Travel Agent         │
    │    (Claude / GPT / Local LLM)   │
    └──────┬──────┬──────┬──────┬─────┘
           │      │      │      │
    ┌──────▼──┐ ┌─▼────┐ ┌▼─────┐ ┌▼───────┐
    │Kiwi.com │ │Triv- │ │Yelp  │ │Weather │
    │  MCP    │ │ago   │ │MCP   │ │  MCP   │
    │(flights)│ │MCP   │ │(food)│ │        │
    └─────────┘ │(hotel)│ └──────┘ └────────┘
                └──────┘

    Agent orchestrates:
    1. Search flights → Kiwi.com (virtual interlining)
    2. Search hotels → Trivago (price comparison)
    3. Search restaurants → Yelp (tapas near La Rambla)
    4. Check weather → June forecast for Barcelona
    5. Synthesize → Present unified trip recommendation

This pattern demonstrates the core value proposition of MCP for travel: a single AI agent can search across multiple travel services simultaneously, synthesizing results into a coherent recommendation. The hosted endpoints (Kiwi.com, Trivago) make this pattern particularly accessible since no self-hosting is required.

Pattern 2: Itinerary Builder with Ground Truth

    ┌─────────────────────────────────┐
    │       Itinerary Builder Agent   │
    └──┬──────┬──────┬──────┬─────┬───┘
       │      │      │      │     │
  ┌────▼──┐ ┌─▼────┐ ┌▼────┐ ┌▼──┐ ┌▼───────┐
  │Mapbox │ │Airbnb│ │Trip-│ │DL │ │Currency│
  │  MCP  │ │ MCP  │ │Adv. │ │MCP│ │  MCP   │
  │(route)│ │(stay)│ │(POI)│ │   │ │        │
  └───────┘ └──────┘ └─────┘ └───┘ └────────┘

  Day-by-Day Plan:
  ┌──────────────────────────────────┐
  │ Day 1: Arrive Barcelona         │
  │  - Airbnb in El Born (€95/night)│
  │  - Walk 12 min to La Boqueria   │
  │  - Dinner: Cal Pep (4.5★ Yelp)  │
  │  - Weather: 26°C, sunny         │
  │  - Budget: €95 + €45 = €140     │
  │         (= ¥22,400 JPY)         │
  ├──────────────────────────────────┤
  │ Day 2: Montjuïc + Beach         │
  │  - Route: 2.3km walk via Gothic │
  │  - Translated menu assistance   │
  │  ...                            │
  └──────────────────────────────────┘

The itinerary builder pattern combines accommodation search, routing/navigation, reviews, translation, and currency conversion into day-by-day plans with real geographic and pricing data. Mapbox’s route optimization and isochrone tools are particularly valuable here — calculating walking times between attractions and reachability areas from the accommodation.

Pattern 3: Real-Time Travel Monitor

  ┌───────────────────────────────────┐
  │     Travel Monitoring Agent       │
  │   (runs continuously or on cron)  │
  └──┬──────┬──────┬──────┬───────────┘
     │      │      │      │
  ┌──▼────┐ ┌▼────┐ ┌▼────┐ ┌▼───────┐
  │FR24   │ │NS   │ │Wea- │ │Ferry-  │
  │  MCP  │ │Rail │ │ther │ │hopper  │
  │(track)│ │MCP  │ │MCP  │ │  MCP   │
  └───────┘ └─────┘ └─────┘ └────────┘

  Monitoring Loop:
  ┌─────────────────────────────────┐
  │ ⚠ Flight BA2490 delayed 45 min │
  │   → Recalculating connections   │
  │                                 │
  │ ✓ Train to airport on schedule  │
  │   → NS: IC Direct 09:15 OK     │
  │                                 │
  │ ⚠ Weather alert: Rain 14:00    │
  │   → Suggesting indoor activities│
  │                                 │
  │ ✓ Ferry Piraeus→Mykonos: normal │
  └─────────────────────────────────┘

This pattern uses aviation tracking (Flightradar24), rail disruption monitoring (NS Dutch Railways), weather alerts, and ferry status to provide real-time travel situation awareness. The agent monitors conditions and proactively alerts the traveler to disruptions, delays, or weather changes that affect their plans.

Pattern 4: Award Travel Optimizer

  ┌───────────────────────────────────┐
  │     Award Travel Optimizer        │
  │  "Find best use of my 150K miles" │
  └──┬──────┬──────┬──────────────────┘
     │      │      │
  ┌──▼─────┐ ┌▼───┐ ┌▼──────────┐
  │seats.  │ │Kiwi│ │flights-  │
  │aero   │ │MCP │ │mcp       │
  │(award) │ │    │ │(cash)    │
  └────────┘ └────┘ └──────────┘

  Analysis:
  ┌─────────────────────────────────┐
  │ Route: NYC → Tokyo (June)       │
  │                                 │
  │ Award Options (seats.aero):     │
  │  ANA F: 110K miles + $86       │
  │  JAL J: 80K miles + $130       │
  │  UA J:  70K miles + $5.60      │
  │                                 │
  │ Cash Comparison (Kiwi/Duffel): │
  │  Economy: $847 roundtrip        │
  │  Business: $4,200 roundtrip     │
  │                                 │
  │ Recommendation: ANA First Class │
  │  at 110K = $0.052/mile value    │
  │  (4.2x better than cash econ)   │
  └─────────────────────────────────┘

This niche pattern leverages seats.aero’s award availability data across 25+ mileage programs to help frequent travelers maximize their miles. By comparing award availability against cash prices from flight search servers, the agent can calculate cents-per-mile values and recommend optimal redemptions.

Official and Hosted MCP Servers

Travel stands out among MCP verticals for its high number of official, company-released MCP servers. This section summarizes all 12 identified first-party implementations.

Company Type Endpoint/Repo API Key Required?
Expedia Group Self-host ExpediaGroup/expedia-travel-recommendations-mcp Yes
Mapbox Hosted mcp.mapbox.com/mcp No (free tier)
Trivago Hosted mcp.trivago.com/mcp No
Kiwi.com Hosted mcp.kiwi.com No
Skiplagged Hosted mcp.skiplagged.com/mcp No
Ferryhopper Hosted mcp.ferryhopper.com/mcp No
Flightradar24 Self-host Flightradar24/fr24api-mcp Yes
Variflight Self-host variflight/variflight-mcp Yes
Yelp Self-host Yelp/yelp-mcp Yes
DeepL Self-host DeepLcom/deepl-mcp-server Yes
WeatherXM Self-host WeatherXM/weatherxm-pro-mcp Yes
LiteAPI Hosted mcp.liteapi.travel/api/mcp Yes

The concentration of free, no-API-key hosted endpoints is notable. A travel AI agent can connect to Skiplagged (flights), Kiwi.com (flights with virtual interlining), Trivago (hotels), Ferryhopper (ferries), and Mapbox (maps) without any API keys or self-hosting — a remarkably low barrier to entry for a functional travel search agent.

Ecosystem Gaps

Despite strong coverage in flights, maps, and weather, the travel MCP ecosystem has significant gaps in core travel infrastructure.

Major OTAs and Booking Platforms

  • Booking.com — The world’s largest hotel booking platform has no official MCP server. Only one community wrapper exists (16 stars)
  • Google Flights — No official MCP; community servers use SerpAPI as a workaround
  • Hotels.com — Zero MCP servers of any kind
  • Kayak — Zero MCP servers
  • Priceline — Zero MCP servers
  • Hopper — Zero MCP servers (despite being AI-focused)
  • Trip.com / Ctrip — Zero MCP servers for the largest Asian OTA
  • Agoda — Only a review scraper (2 stars); no booking capability

Ground Transportation

  • Uber — Zero MCP servers
  • Lyft — Zero MCP servers
  • Hertz / Enterprise / Sixt — Zero official car rental MCP servers
  • Waze — Zero MCP servers
  • Google Maps — No official MCP (strong community alternatives exist)

Cruise Industry

The entire cruise industry — Royal Caribbean, Carnival, NCL, MSC, Viking, Celebrity — has zero MCP presence. No search, no booking, no itinerary servers exist.

Activities and Experiences

  • Viator — Zero MCP servers
  • GetYourGuide — Zero MCP servers
  • Klook — Zero MCP servers
  • Musement — Zero MCP servers

Restaurant Reservations

  • OpenTable — Zero MCP servers (Yelp’s official server includes some reservation capability, but OpenTable’s inventory is unrepresented)
  • Resy — Zero MCP servers
  • TheFork — Zero MCP servers

Travel Infrastructure

  • GDS Systems (Travelport, Sabre) — Zero official MCP servers. These systems power the majority of airline and hotel booking infrastructure globally
  • Travel Insurance — Zero MCP servers for any travel insurance provider
  • Visa/Immigration — Zero MCP servers for visa requirements or application tracking
  • Airport Services — No lounge access, parking, or ground transport MCP servers
  • Travel Expense Management — No Concur, Navan, or SAP Concur MCP servers
  • Loyalty Programs — No direct loyalty program management servers (seats.aero covers award search but not account management)

Rail Outside Europe

  • Amtrak — Zero MCP servers
  • Trainline — Zero MCP servers
  • JR (Japan Rail) — Zero MCP servers
  • Indian Railways — Zero MCP servers

Rail MCP coverage is limited to Dutch Railways (NS) and Spanish Renfe — both European-only.

Regulatory and Ethical Considerations

Travel AI agents face specific regulatory and ethical challenges that MCP server developers and users should consider.

Price display regulations. Many jurisdictions require all-inclusive pricing for flights and hotels. AI agents presenting travel prices must ensure they display total prices including taxes and fees, not base fares only. The EU’s Package Travel Directive and the US DOT’s fare advertising rules impose specific requirements on how travel prices are presented.

Booking liability and consumer protection. When an AI agent executes a booking, questions arise about liability for errors — wrong dates, wrong airports, non-refundable bookings made without explicit confirmation. MCP servers that support booking transactions (like LiteAPI and Yelp reservations) need careful confirmation workflows to protect consumers.

Data privacy across borders. Travel inherently involves cross-border data flows. Passenger data (PNR records), passport information, and travel histories are sensitive personal data subject to GDPR, CCPA, and various national data protection laws. MCP servers handling travel bookings should consider what personal data flows through their tool interfaces.

Discrimination in travel recommendations. AI travel agents must avoid discriminatory recommendations based on protected characteristics. Dynamic pricing, which is common in travel, can inadvertently create discriminatory outcomes. The EU AI Act classifies certain AI systems in travel as high-risk if they affect access to services.

Web scraping vs. API access. Several travel MCP servers (particularly the Playwright-based airline and hotel servers) work by automating web browsers rather than using official APIs. This approach may violate terms of service and raises questions about rate limiting, accuracy, and reliability. Official API-based servers (like the hosted endpoints from Kiwi.com and Trivago) provide more reliable and legally sound access.

Environmental transparency. With growing attention to travel’s environmental impact, AI travel agents may face expectations to disclose carbon emissions for recommended itineraries. The timcp server (Travel Impact Model) addresses this for flights, but comprehensive environmental impact data across hotels, ground transport, and activities remains unavailable through MCP.

Getting Started

For Solo Travelers

Start with the free hosted endpoints that require no API keys: connect Kiwi.com MCP (flights with creative routing), Trivago MCP (hotel comparison), and Mapbox MCP (maps and navigation). Add a weather MCP server for destination conditions. This gives you a functional AI travel search assistant with zero configuration cost.

For Travel Agents and Advisors

Combine flight search (flights-mcp with Duffel API for 300+ airlines), hotel comparison (Trivago hosted), vacation rentals (mcp-server-airbnb), reviews (tripadvisor-mcp), and currency conversion. For clients with loyalty programs, add seats.aero for award availability analysis. This creates an AI assistant that can research options across the travel landscape while you focus on client relationships and complex bookings.

For Travel Tech Developers

Use the official hosted endpoints as building blocks for travel applications. The free hosted MCP servers from Kiwi.com, Trivago, Skiplagged, Ferryhopper, and Mapbox provide immediate access to flight, hotel, ferry, and mapping data. LiteAPI adds booking capability. Build orchestration layers that coordinate across these services for specific travel use cases.

For Frequent Flyers and Travel Hackers

The travel-hacking-toolkit bundles 6 MCP servers specifically designed for finding deals — including Skiplagged (hidden city fares), Kiwi.com (virtual interlining), and seats.aero integration for award availability. Combined with currency conversion, this creates an AI-powered deal-finding agent.

For Property Managers

Lodgify and Guesty MCP servers (38 tools) connect AI agents to vacation rental management platforms. Combined with Airbnb search for competitive analysis, weather data for seasonal pricing, and maps for location-based guest recommendations, these tools enable AI-assisted property management workflows.

For Corporate Travel Managers

While corporate travel management (Concur, Navan) lacks MCP servers, you can build monitoring workflows with Flightradar24 (flight tracking), NS Rail (European train status), weather alerts, and currency conversion. The timcp server adds carbon emission tracking for sustainability reporting on corporate travel.

Conclusion

Travel and hospitality represent one of the most commercially active MCP verticals, with 12 official first-party servers — more than most industry categories. The free hosted endpoints from Kiwi.com, Trivago, Skiplagged, Ferryhopper, and Mapbox create an unusually accessible entry point: a functional travel search agent can be built with zero API keys and zero self-hosting.

The ecosystem’s strengths cluster around flight search (10+ implementations), maps/navigation (Mapbox official + strong community Google Maps servers), weather (multiple options including API-key-free Open-Meteo), and vacation rentals (mcp-server-airbnb at 406 stars). Aviation tracking benefits from two official servers (Flightradar24 and Variflight).

The gaps are equally clear. Major OTAs (Booking.com, Kayak, Priceline), ride-sharing (Uber, Lyft), the entire cruise industry, activity platforms (Viator, GetYourGuide), and core travel infrastructure (GDS systems, travel insurance, visa services) have no MCP presence. Ground transportation beyond maps is nearly absent. Rail coverage is European-only.

For the travel industry, the trajectory is clear: companies that release official MCP servers (like Expedia, Trivago, and Kiwi.com) position themselves for AI agent distribution — a channel that could become as significant as mobile apps and metasearch. Companies without MCP presence risk being invisible to the growing population of AI-assisted travelers. The hosted, free-to-use model pioneered by several travel MCP providers may become the standard approach for travel platforms seeking AI agent adoption.