The poster tried to have Claude fully book flights and hotel via Expedia and Kiwi connectors, found them poor, and ended up building a travel booking project instead.
The poster tried to have Claude fully book flights and hotel via Expedia and Kiwi connectors, found them poor, and ended up building a travel booking project instead.
The poster tried to make the default Claude mobile/web app fully book a Warsaw-to-London trip, using Expedia and Kiwi through Claude’s connectors. Claude could search options, but the attempt dragged on because the booking flow, especially taking payment safely without exposing card details to the model or connector, was not solved; the poster then turned the experience into building a travel-booking product instead.
The trip itself was not fully booked by Claude, and the process stretched into a six-month project. In the end, the poster says a later model release changed what felt possible, but the original Expedia and Kiwi connector attempt was poor.
The biggest catch was payments: the post frames bank-grade card handling, without the AI or services ever seeing the card data, as the central blocker.
Build me a small travel planning project for a trip from [origin city] to [destination city] on [dates]. It should pull flight and hotel options from [sites or APIs you prefer] into one ranked comparison file using my budget of [amount] and preferences of [preferences]. Stop once the shortlist is ready and show me the top choices with links, and don't book or pay for anything.
Score every demo request against your ICP, update HubSpot and ping the right rep in Slack within a minute of the form submit.
Every new Zendesk ticket gets a category, priority and sentiment tag plus a draft reply in an internal note before an agent opens it.
Copilot and Claude review each PR for bugs, missing tests and risky changes before a human looks, so reviewers can focus on design.