Ask ChatGPT to find flights from London to Tokyo under £800 in April, and it might give you a decent answer. Then ask "what if I fly a week later?" and it forgets everything you just told it - your budget, your origin city, your dates. You are back to square one.
JetSet AI, a new conversational flight search tool, attacks this specific problem. Instead of running each query through a stateless serverless function (the standard approach where each request starts fresh with no memory of previous ones), it keeps each user session running on a persistent virtual machine. Your full conversation state stays in memory, so follow-up questions actually work.
The tool pulls real-time pricing from live flight APIs and generates direct booking links to Skyscanner and Kayak with your search pre-filled. It is free to use without creating an account.
The limitations are honest and clearly stated: there is no cross-session memory (close the tab and your context is gone), no hotel or car rental search, and no automatic booking. It finds and compares flights through natural conversation, then hands you off to an existing booking platform.
Persistent VMs cost more to run than serverless functions, which is why most AI tools do not use them. But for multi-turn transactional searches where every follow-up depends on remembering what came before, the tradeoff makes sense. It is a narrow tool solving a narrow problem, and that focus is what makes it interesting.