https://traveldna.pages.dev/ Every travel app asks the same weak question: "What do you like?" Food, museums, adventure, nightlife. Most people click a few boxes and move on. But that data is shallow, static, and usually wrong. The better answer is already sitting in your inbox. Years of bookings, confirmations, tours, stays, tickets, trains, flights, and receipts tell a much clearer story about how you actually travel. We built a travel preference inference engine. The user connects Gmail, chooses a timeframe, and the system extracts travel evidence from their booking history. It detects where they have been, what platforms they use, and what kinds of experiences show up repeatedly: food tours, architecture walks, train travel, boutique stays, surf lessons, wine tastings, museums, nightlife, and more. That becomes a structured "Travel DNA" profile: categories, keywords, counts, confidence, evidence, and an embedding that can be used by downstream planning agents.