The Gondola MCP: How AI Agents Use Gondola
For a few years now, Gondola has been building one engine for every part of hotel rewards. We search every major chain at once and surface the same Member rates a hotel would show you logged in on its own site. We keep track of your stays after you book and rebook automatically when the rate drops, which saves the average Gondola traveler $413 a year. And we track your loyalty balances and credit card points, which has surfaced more than $3,100 worth that the average member didn't know they had.
The Gondola MCP is how AI agents get to use that engine. It's a free MCP server you add to Claude, ChatGPT, Cursor, Windsurf, Claude Code, or any other MCP-compatible client. Once it's added, the agent can do everything Gondola does on the hotel and rental car side, right inside the chat.
With the MCP connected, an agent can search the same Member rates a traveler would see logged in. It can redeem their actual loyalty points instead of paying cash. Bookings go direct to the chain, so every status night and category bonus is preserved. And the agent keeps track of the rate after the booking, rebooking automatically if the price drops. None of these are things a public travel API surfaces, which is why most AI travel agents today still hand back rack rates and OTA links.
What the Gondola MCP does
Once a traveler adds the Gondola MCP to their AI client, the agent can:
- Search every major hotel chain in one query: Marriott, Hilton, Hyatt, IHG, Accor, Wyndham, and the long tail of independents
- Return the same Member rates, AAA rates, and package rates the hotel shows a logged-in member, often 10 to 25% below the public price its travel API returns
- Compare cash and points in one response with the cents-per-point math done
- Search rental cars across the major suppliers and show which of the traveler's credit cards cover insurance
- Pull the traveler's loyalty balances, card portfolio, preferences, and trip history once they authorize the connection
- Book direct with the chain so every loyalty point, status night, suite upgrade, and credit card category bonus is preserved
- Keep track of the rate after a booking and capture the savings if the price drops, the same way Gondola already does for travelers on gondola.ai
The agent handles natural-language requests like these:
- "Find me a Hyatt in Tokyo for 3 nights in October under 25,000 points"
- "Compare cash vs. points for the Waldorf Astoria Maldives, December 20-27"
- "Which Marriott in Paris gives the best value if I use my 80,000 Bonvoy points?"
- "I need a hotel near LAX for a layover tonight, cheapest member rate"
No special syntax. The agent maps the request onto the right MCP tools and Gondola fills in the rest.
What's in the MCP
The Gondola MCP has about 30 tools, covering hotels, rental cars, loyalty, and rate monitoring. Public hotel search works without any account. Tools that touch a traveler's data require a one-time OAuth login through gondola.ai, which the traveler can revoke anytime from their Connected Agents page.
| Domain | Tools |
|---|---|
| Hotels | search_hotels, get_hotel_details, get_multi_night_rates, compare_rates, get_similar_hotels, get_hotel_reviews, get_hotel_stats, predict_price, book_hotel, get_booking_link |
| Rental cars | search_vehicles, get_vehicle_details, book_vehicle, cancel_vehicle_booking, get_vehicle_booking_link, get_vehicle_booking_coverage, credit_card_coverage |
| Loyalty and traveler context | get_loyalty_accounts, optimize_loyalty_portfolio, get_payment_methods, get_travel_profiles, get_upcoming_trips, get_past_trips |
| Rate monitoring | get_rate_alerts, create_rate_alert, delete_rate_alert, diagnose_rates, get_suggested_searches |
How to install
Setup is one configuration block in your client. Copy-paste instructions for Claude Desktop, Claude Code, ChatGPT, Cursor, and Windsurf live at gondola.ai/mcp. The server endpoint is https://mcp.gondola.ai/mcp.
Public search works without any account. The authenticated tools (loyalty balances, booking, rate alerts, trip history) use OAuth. A traveler authorizes the agent once on gondola.ai, and they can revoke that access anytime from their Connected Agents page.
The Gondola MCP is free, and Gondola doesn't take a cut of bookings. Every booking goes direct to the hotel.
FAQ
What is an MCP?
MCP (Model Context Protocol) is an open standard that lets AI assistants like Claude and ChatGPT connect to external tools and data sources. Adding the Gondola MCP gives your agent the ability to search real hotel availability, compare rates, and book hotels.
Is the Gondola MCP free?
Yes, the Gondola MCP is completely free. No account, API key, or credit card required. Install it and start searching immediately.
What hotel chains are supported?
The Gondola MCP searches Marriott, Hilton, Hyatt, IHG, Accor, Wyndham, and more, every major hotel chain worldwide. You get Member rates, AAA rates, package rates, and loyalty points redemptions in every search.
What rates does the MCP return?
Unlike OTA APIs that only return public rack rates, the Gondola MCP returns Member rates, AAA rates, and package rates, the same pricing you'd see logged in on the hotel's own website. These rates are only available by booking directly with the hotel.
Can my agent book hotels through the MCP?
Yes. The MCP includes a book_hotel tool and returns direct booking links for every result. Bookings go directly through the hotel, so travelers earn their full loyalty points, status nights, and credit card rewards.
What AI clients work with the Gondola MCP?
Any MCP-compatible client works, including Claude Desktop, Claude Code, ChatGPT, Cursor, Windsurf, and any custom AI agent that supports the Model Context Protocol.
Use Gondola from any AI agent. Get started at gondola.ai/mcp.