Original Summary

I travel internationally once or twice every single year. It’s a huge part of life for my wife and me. We’ve been to 9 countries so far, but there has always been one specific thing that drained a ton of energy from me: selecting a great hotel. At first, I used traditional tour agency deals, but stopped once I realized it was mostly a waste of money. Then I tried finding hotels physically after landing, yk walking around with luggage and booking whichever looked best. Yeah this was one of the most stupidest thing I've ever did. Anyways, I did that for 2 or 3 trips until my wife was completely fed up and told me: “I don't care if you spend an extra $500, but I am not dragging luggage to 10 different hotels after a 12 hour flight” After a lot of skepticism I switched to booking online through OTAs. Over the last few years I've booked through Booking(.)com most of the hotels. But I noticed that I developed a very specific (and ridiculously time consuming) habit. I’d pick 10 candidate hotels on Booking, and over the course of literally 3 days, I’d check pretty much like100 reviews for each one to get a better grasp on which one would fit us the best. And it matters a lot the us because my wife always needs a bathtub to relax, and without one she's always complaining before and after showering. Plus a noisy or weak AC completely ruins her mood for the day, so I had always to be extremely careful with my choices. Even then, I sometimes failed to get the best hotel option because OTAs frequently hide or downgrade negative reviews, so you still get surprised by bad AC or paper thin walls. Eventually, I started copy pasting entire pages of comments into ChatGPT to summarize them. It helped, but copying 100+ comments across multiple tabs was still exhausting. So over the last 6 months, I decided to solve this properly. I work in engineering, so I’ve been building an open source travel system (mostly focused on flight aggregation because hotels seemed like a worse story, but actually it was the other way around. Anyway a month ago I finally finished the testing period and added the hotel engine to the tool, and even tested it for our upcoming trip to Italy in January). It connects directly as an MCP connector for LLMs like Claude, chatgp etc. Instead of pulling superficial data from OTA listings, it connects into live wholesale pricing data which I managed to get access to (which is quite often cheaper than public OTA rates) and checks recent reviews from across the entire web simultaneously. I tested it for the Italy trip and Claude navigating the tool read close to 1,000 reviews and it was really helpful. I used this prompt: "Find a hotel in Italy with a great central location, but strictly filter out any place where recent reviews mention noise, weak AC, or lack of a bathtub." And obviously, it gave me offers that match that. I made some parts of it open source, but overally the whole thing is free to search. If you want to plug

  • 情报分类：技术学习与提效
  • 分类依据：内容涉及技术、AI、软件工具或工程实践
  • 信息来源：Reddit · SideProject
  • 发布时间：2026/10/11 03:57:39