Three Weeks From Annoyance to SaaS: How Claude Wrote 90% of the Code for a Flight-Search Service

The best products are born from personal pain. The line has long since become a startup cliché, but occasionally a case comes along that restores its original meaning. A developer from Russia turned his frustration with searching for flights into a full-fledged SaaS service in three weeks, and the Claude language model wrote ninety percent of the code.

The original task sounded mundane: find tickets to Bali for five people, with flexible departure dates and a set trip length. Anyone who has tried something similar knows the feeling: aggregators like Aviasales or Skyscanner work great when you fly alone and know your exact dates. But add a few variables, such as a group of travelers, a floating date range and a trip-length requirement, and the interface starts to stall. You end up checking dozens of combinations by hand, tab after tab. That was the moment something clicked for the project's author: instead of spending hours on manual search, he decided to automate the process.

The result is a Telegram bot that can check more than a thousand route combinations an hour and find options with savings of up to fifty-two percent compared with a standard search. One of the key insights uncovered along the way: a round-trip ticket can turn out to be forty percent cheaper than the sum of two one-way tickets. It seems obvious, but aggregators far from always show this difference clearly, especially on complex routes with connections.

The technical side of the project deserves separate attention. The author deliberately chose an "AI writes all the code" approach and entrusted most of the development to Claude. By his estimate, the model handled ninety percent of the tasks, from backend logic to DevOps configuration. This is not just boilerplate generation: it is a complex architecture that includes reverse-engineering the closed APIs of airlines and aggregators, bypassing anti-bot systems, and building analytics infrastructure with CTR and retention metrics and conversion funnels. Three weeks from zero to a working product with real users is a pace that a couple of years ago would have required a small team and a couple of months of work.

That said, a caveat is in order: the case also raises uncomfortable questions. Reverse-engineering other companies' APIs and bypassing protection mechanisms is a practice that sits on the edge of legality. Most airline aggregators explicitly prohibit automated data collection in their terms of use. That doesn't mean such projects have no right to exist, but scaling a service like this will inevitably run into legal and technical barriers. Companies regularly update their defenses, and case law on scraping remains contradictory.

The more interesting aspect here is something else: the speed with which language models let a single person build products at a level once available only to teams. When Claude takes over the routine of writing code, the developer can focus on product thinking: what exactly to build, for whom, and how to measure success. The author stresses that he was thinking about backend, DevOps, UX and business metrics at the same time, and that this ability to hold the whole picture in mind is what separates a successful project from a technical demo.

The case fits into the growing "one-person SaaS" trend, in which AI tools act as a multiplier of what individual developers can do. We see more and more stories of one person with a clear understanding of a problem and access to modern language models shipping a product faster than a corporate team gets through an approval cycle. This does not remove the need for teams on large-scale projects, but it radically lowers the barrier to testing hypotheses.

Three weeks, one developer, a language model as co-author and a live service at the end. A formula that looked exotic a year ago is becoming the new normal. The question is no longer whether AI can help write code, but whether the developer has enough product sense to steer that power in the right direction.