A flight search engine that combines flights from different airlines? (2014)

It's called "virtual interlining".

The startup I co-founded (Adioso - YC W09 [1]) tried to do it, including having discussions with a travel insurance provider about offering "layover protection" - so that if one leg was delayed causing you to miss your onward leg(s), your costs are covered. Kiwi.com does this now.

We worked on it from about 2008 till 2013 then basically gave up, as it was too hard to offer a service that customers could really love and trust. (It wasn't for nought; the technology we developed was valuable, and the company was able to rebrand and pivot and now does important work for airlines to optimise loads and fares [2], though I left when the rebrand/pivot happened).

The thing that makes it hard to do is it's basically impossible to get all the flight inventory, including fares and seat availability, that's complete and up-to-date enough to deliver a service that customers can trust.

The engineering challenge is one thing - solving a multi-dimensional travelling salesman problem (price and distance/duration) highly repetitively - but you can solve that with enough smart engineers and "compute", which ITA did in the early 2000s, and on a smaller scale, our team did a decade later.

But you could build the most beautiful routing engine the universe has ever seen, and still have a user experience that's kind-of garbage because the industry just keeps the flight inventory data so locked down.

These days there are APIs and feeds available from the major distribution platforms - Sabre, Amadeus and Travelport, but it's still not comprehensive. You often still need to negotiate individual agreements with major airlines in order to be able to publish and sell their fares. And even then, many of the low-cost airlines (which are often most of interest to travellers who want to find the cheapest route and deal with self-transfer) are not available through these distributors, and some, like Southwest, have blanket refused to be on 3rd party search sites, only starting to relax that position very recently and only with the dominant platform [3]. Kiwi.com has only recently come to a partnership agreement with Ryanair [4] after being in legal battle with them for years [5]. (I hate the thought of having to be at war with your most important partners).

Others have mentioned Skyscanner, which was always the closest to us in what we were each trying to offer (we talked briefly with them about being acquired by them).

Right from the beginning when we got funded for Adioso, my mind became fixated on the thought "if only you get every single flight in the world loaded into one big graph database, what could you do with it?", but it turned out to be a very big "if".

[1] https://techcrunch.com/2010/08/31/adioso/

[2] https://amadeus.com/en/blog/articles/creating-a-private-resa...

[3] https://www.forbes.com/sites/geoffwhitmore/2024/05/25/southw...

[4] https://media.kiwi.com/company-news/kiwi-com-and-ryanair-ann...

[5] https://www.travolution.com/news/kiwi.com-celebrates-three-w...

Happy to see ITA software mentioned in the comments: https://matrix.itasoftware.com/search

I still have fond memories of their legendary (pre-leetcode) coding challenges [0] posted on the T (they also hosted the Boston Lisp users group in the early 2000s which was filled with mind-blowingly incredibly brilliant people, everyone there seemed to be an expert in software and had a PhD in some other, non-related field)

Having worked a bit in the travel industry, I highly recommend that you never book through a third party (by all means use their search). Third party apps are not allowed by airlines to charge less than the airline and typically have abysmal customer service, and I can assure you any "add-ons" offered by a third party are ultimately a scam.

0. https://github.com/mattbraz/ita-puzzles

From the great presentation linked: http://www.demarcken.org/carl/papers/ITA-software-travel-com...

> Amazingly, the graph diameter is often as high as 20: there are airports that can take 20 flights minimum to get between

I wonder if that's still true. It's hard to imagine. And just the thought of having to optimize that search function made my cortisol levels spike a bit.

Never book anywhere else but the airline website. It's not worth it, these days. There might be some deals but when and no longer if, IRROPS happens you will face an uphill battle to fix your problem playing a triangle between the airline and the OTA you booked with. Flying has changed: past covid and Boeing-being-shit-revealed there are not enough long range planes, in Europe because of the Ukrainian war Hungarian ATC is overwhelmed which creates ripple effects all over and nothing is on time. And that's just the regions I know about. https://travel.stackexchange.com/q/174173 has more.

One search feature I wanted was a way to look up a list of all flights leaving an airport on a certain day.

In the early stages of vacation planning, it’s be fun to see a list of all possible direct flights to evaluate my options, but the use case of doing flight searches with an unknown destination isn’t too common. Basically, i want to be able to browse flights like a bus schedule and just see what the possibilities are from a particular start point

tomhoward is correct. It's called "virtual interlining". I started the team that built it at Hopper (started working on it in Fall 2020 and we launched our first version in Fall 2021). They're called Mix & Match fares on the Hopper app. It's a really hard problem. So hard that we didn't even try to solve the "best" flight. We used a bunch of heuristics to find what we believed would be "better" flights. We measured what we called "beat rate" which was the rate that a VI flight was the cheapest on our flight list. When I left the team, our beat rate was around 0.5 (so a VI flight was the cheapest half the time).

Adjacent plug for a free Chrome Extension: FlyOnTime embeds flight delay probability straight into Google Flights. Very useful when you're evaluating a number of flights from the same origin and destination but at different times.

Helps you really know what you're paying for.

https://chromewebstore.google.com/detail/flyontime/mjhocoppe...

One use case I’d love to see is the ability to find the best price for multiple city tickets.

I’m fairly flexible on dates, but getting the best price for it is a super iterative process. Having some similar functionality like excel’s solve function would be awesome to find optimal dates within a range for each destination.

In my experience, these tools have lost their value in the post-COVID world because the airlines have (a) collectively hiked their prices, and (b) have slashed and otherwise aggressively changed their route structure.

I spent many a happy hour <2020 on ITA Matrix. It was well worth learning as a tool back then despite its terse interface and tricky syntax.

But these days? The sorts of clever tricks you could do before are no longer viable for the reasons I outlined at the top, i.e. connections no longer exist and/or the price is no longer attractive.

And before anyone tries to tell me otherwise, my experience is bang up to date. What I said above is based on 2024 experience.

I have long wanted to build an engine that let's you book a vacation on points. You connect all the points programs you have and it shows you how to leverage them to get the best deals on flights and hotels. The trick here is that you can transfer points from credit cards to some airlines and hotels, and sometimes the points are worth more when transferred. But, integrating with all these systems is the problem, and it's a big problem. Especially since there is no way any of the players want this to happen. Maybe it could be built as a browser plugin?

I've never understood the comments on all of the posts like this advising people to never book via a third party. Is it an American thing?

I've probably been on around 250 flights in the past 7 years in Europe and I can't remember one time that it was cheaper to book directly than through some third party. Sometimes it's the same price to book direct, but it doesn't matter either way: the support you're going to get is the same, and the insurance or whatever depends on... exactly whatever you've paid for already.

The leading response does a good job calling out that this is pretty risky for the traveller. Similar to skip lagging, if anything goes wrong you can end up is a pretty difficult position.

Interesting to see Hipmunk [1] among the suggestions. I believe Steve Huffman (reddit founder) started that and was yc funded.

[1] https://travel.stackexchange.com/a/26833/100794

Google built something amazing in Google Flights [1] (and they haven't ruined it yet).

It's no-nonsense, easy to filter by number of stops, and 'date grid' is great for scoping out savings by departing a day or two earlier/later without having open multiple tabs as other sites necessitate.

Only criticisms are extremely mild ones: it defaults to 'return', doesn't remember your currency, and bizarrely defaults to a month ahead for the departure date (actually, they must have very recently fixed this because it doesn't do that anymore!)

[1] https://www.google.com/travel/flights

I think Google Flights can do that?

I want to search from multiple cities. I have often saved a lot of money by flying out of an airport 1.5 hours drive away from home.

I worked on this for tripstack a while ago. They offer it through their partners. Kiwi is the other one who made it a thing. Now it’s more common to see it all over.

Skyscanner

Aviato

Isn't that what Expedia does?

kiwi.com is also pretty good at this.

skiplagged shows you “self-transfer” flight options, which i think solves what OOP was asking

Kayak.com might be what you are looking for.

A web search engine that combines results from different web sites? (2022) (openai.com)