If you’re reading this, you probably already know that “good single-player airline simulator” is a phrase that doesn’t really land anywhere. There’s no obvious answer when someone asks. There’s AirlineSim, which is the canonical depth standard, but it’s a browser MMO that demands daily logins and never sleeps. Airline Tycoon is twenty-five years old. World of Airports is a ground-handling game. Air Manager is something else entirely. The simulators with the depth you want require you to live in them. The single-player games with the convenience you want skip the depth.
We’ve been building Skyline for the better part of a year because we wanted the version of this that doesn’t exist. A serious airline management simulator that runs entirely on your machine, has no daily-grind expectations, and lets you actually think about the decisions a real airline CEO makes across the kind of timeframes those decisions actually play out on.
This is the first devlog. We’re going to use this space to talk about what we’re building, how we’re building it, what we’ve learned in the process, and where the project is going. We’ll also be honest about what hasn’t worked, what we got wrong, and what we changed our minds about. Aviation enthusiasts deserve that kind of post-mortem more than most audiences, because you’ll catch us if we lie.
So: what is Skyline, why does it exist, and where are we today.
What Skyline is
Skyline is a single-player airline management simulator built around real aviation economics. You run an airline. Pick the era, pick the hub, pick the strategy, then live with the consequences in a world that doesn’t pause for you to catch up. Fleet acquisition, route planning, pricing, schedule design, hub operations, certification, financing, competition with AI carriers that have their own strategies. The whole thing.
The simulation is grounded in actual aviation data. Passenger demand calibration comes from the US Department of Transportation’s T-100 segment data and Eurostat’s avia_par filings. Aircraft performance follows BADA methodology, the same fuel-burn and operating-cost framework used in real route planning software. The economic model is fit against years of historical operating data across different market conditions. None of this is unusual for the genre’s most serious entries. What is unusual is doing it without a multiplayer server attached.
You pick the year you want to start in, anywhere from the early 1990s onward, and the simulation gives you that year’s fuel prices, passenger demand patterns, aircraft availability, regulatory environment, and competitive landscape. Start a US carrier in 1995, in the middle of the post-deregulation consolidation wave. Start a European LCC in 2003, the year Ryanair turned profitable. Start a Polish flag carrier in 2017, navigating consolidation and slot pressure at WAW. Each starting position has its own constraints and its own historical context, and the world evolves from there. Including events that genuinely happened, like the post-2008 demand collapse, the 2014 oil price drop, the 2020 pandemic.
The simulation is offline-first and pauseable. You can play for thirty minutes on a Tuesday night and not log in again for a month, and your airline waits exactly where you left it. There are no decay mechanics, no daily login penalties, no scheduled events you’ll miss if you don’t show up. If you want to study a single route’s economics for an hour while the game is paused, you can do that. If you want to advance ten years in a few minutes to see how a long-term strategy plays out, you can do that too.
That’s what we’re building.
Why this is hard, and why most attempts fail
A single-player airline simulator with real depth is one of those project ideas that looks easy from the outside and turns out to require a surprising amount of infrastructure to do honestly.
The reason most attempts go shallow is that airline economics doesn’t simplify well. You can’t model a route as a single number, because the same route flown at different frequencies, with different aircraft, against different competitors, at different times of day, in different seasons, produces wildly different outcomes. You can’t model fleet decisions without modeling several things at once: the negotiated price of an aircraft against the list price (typically a 35-55% discount, depending on the carrier’s relationship with the manufacturer and the order size), the configuration choices that lock at different milestones in production, the financing structure (cash, finance lease, operating lease, sale-leaseback, EETC, export credit, each with meaningfully different P&L consequences), and the maintenance reserves that will eventually come due. You can’t model demand without modeling segmentation, since business travelers and leisure travelers respond to entirely different signals. Or seasonality. Or day-of-week patterns. Or the way an LCC entering a market induces new demand that wasn’t there before.
You can simulate all of this. You just need to actually build it.
Most games skip these layers and substitute a single profitability number per route, a single price per aircraft, a single demand curve per city pair. The result is a game where every airline plays roughly the same way because the levers that differentiate real airlines have been collapsed into one. Why does Ryanair price the way it does? Why does Emirates fly the routes it flies? Why did Spirit’s strategy work in 2015 and break in 2024? These questions have specific answers grounded in specific economic structures, and if your simulation can’t represent those structures, those answers aren’t available to the player.
The other failure mode is the opposite. You build all of that depth and then make it an MMO. AirlineSim went this route, and it works for the audience it has, but it imposes a daily-grind tax on players who would otherwise love the depth. You can’t pause an MMO. You can’t play strategically across a decade of simulated time if every real day equals one simulated day. The depth is real, but it’s locked behind a time commitment most adults can’t sustain.
We think the gap between these two failure modes is where a good single-player airline simulator lives.
What we’ve built so far
We’ve been at this for about a year now, and a lot of the time has gone into making the foundations right rather than building features on top. The current state of the project, in rough terms:
The simulation engine is a pure-functional, deterministic Rust codebase. About 150 files, a stable public API surface, around 800 tests at last count. Same seed, same outcomes, byte-identical. This is a property aviation enthusiasts will appreciate. You can replay a save and analyze what would have happened differently if you’d made a different decision, with the rest of the world held constant.
The economic core runs the gravity-based demand model, multinomial-logit market share computation, and BADA-methodology fuel burn. The model is calibrated against six years of US DOT T-100 data and Eurostat aviation filings. We’re aware that v1 of the seasonality calibration is hand-tuned to publicly-known industry patterns rather than full residuals-against-T-100, and that’s a known follow-up. We’d rather ship an honest first version and iterate than over-promise.
The AI competitor system is the part that we think genuinely distinguishes Skyline from most of the genre. Eight curated AI airlines operate under the same rules the player does. They expand into profitable markets and retreat from unprofitable ones. They negotiate aircraft orders. They participate in the used-aircraft market. They go bankrupt when they overextend, and new entrants emerge when market conditions reward them. The world is not waiting for you to play.
The airline founding system lets the player customize a founder, pick a jurisdiction (each with its own tax rate, regulatory regime, route rights, and union strength), structure capital through one of five funding paths (founder equity, friends and family, angel, private equity, IPO), choose a brand archetype (legacy, LCC, ULCC, premium, charter, cargo, regional, hybrid), and pick a network strategy. Founder traits, about 40 of them, propagate through downstream simulation, affecting negotiations decades later. The Air Operator Certificate progresses through real stages from pre-application through demonstration flights to issuance, with durations scaled by jurisdiction complexity.
The fleet acquisition system covers the full real-world decision sequence: requesting and optioning production slots, negotiating terms with manufacturers, configuring aircraft (with the engine choice locking at firm-up, the cabin configuration at delivery minus six months, everything fully locked after delivery, matching real production constraints), and financing across all six financing types we mentioned earlier. There’s a used-aircraft market with condition gradings. There’s a lease market with multiple lessor archetypes, each with different risk appetites, pricing premiums, and portfolio focuses. AI competitors participate in all of this, not just the player.
The UI is built in React with a Tauri desktop shell, using a locked dark-mode design system. The dashboard, world map, schedule, fleet, finance, news, and certification screens are functional. The operations control center, which is the screen where you actually run your airline day-to-day, shows aircraft rotations on a Gantt timeline, hub bank structure analysis, real-time KPIs (on-time performance, dispatch reliability, utilization, tight turns, maintenance status), and a network pressure panel that flags slot saturation and delay propagation risk.
We’ve also done the unglamorous infrastructure work: deterministic save/load, scenario configuration, COVID modeled as a regime in the simulation, fuel price time series, FX modeling, exogenous events. None of this is interesting to talk about and all of it has to be right for anything else to work.
What we’re not building
This part is as important as what we are building.
No multiplayer. Not in 1.0, almost certainly not ever. Multiplayer airline simulation is a different game with different design pressures, and the moment you add it, you have to make every economic decision asynchronous, time-bounded, and abuse-resistant. AirlineSim does this well. We don’t think we’d add to what AirlineSim already does. The point of Skyline is to be the best single-player option, full stop.
No real airline names and liveries in the base game. The base game ships with fictional carriers. We made this choice for licensing reasons and we’re standing by it. Real airline names and liveries will be supported through mods, and we’ll provide a clean API for that. The modding community will handle this within days of launch, and we’d rather have that than negotiate hundreds of licensing agreements that don’t add anything to the simulation.
No microtransactions, no DLC fragmentation, no live-service mechanics. Skyline will be paid up-front, somewhere in the $25-35 range on Steam when it ships. Future content will come from updates and mods, not from a season pass. If we add a major expansion someday, it will be a major expansion, not a battle pass.
No gameplay that requires daily attention. If you want to disappear for a month and come back, your airline will be exactly where you left it. The game does not punish you for having a life.
No fake difficulty through arbitrary failure events. Your airline will fail when the simulation says it should fail. When your cash position can’t service your debt. When your operational metrics collapse. When you’ve made strategic decisions the market punishes. It won’t fail because a random event killed you. If you go bankrupt, the reason will be visible in your financial statements and your operating data, and you’ll be able to read it.
Where we are today, honestly
We’re 14 to 18 months from a closed alpha. That estimate is realistic, not aspirational, and we’d rather tell you the truth than commit to a date we can’t hit.
The simulation engine is in good shape. The UI is taking shape. The remaining work, in rough order, is: expanding the demand and revenue models to include the segmentation and seasonality depth we know is missing in v1, polishing the founding and fleet acquisition flows we’ve already shipped, adding the systems we haven’t built yet (some operational depth, some economic dynamics, some longer-term gameplay loops), and doing the calibration work that turns a working simulation into a believable one. After that, we run a closed alpha with a small group of aviation-domain advisors and active community members, iterate on what we find, expand the alpha, and eventually ship.
We’re going to talk about all of this in this devlog as we go. The next post will probably be a deep dive on the demand model: what we calibrated, what we cheated, what the limitations are, what we’re going to fix in the next iteration. After that, probably something on the AI behavior layers, or on how we model fleet financing, or whatever the most interesting thing happens to be when we sit down to write.
We’re also going to post things that aren’t strictly devlogs. Design philosophy pieces about specific choices we made. Technical posts about the engine architecture. Occasional pieces about what we got wrong. We’d rather have a small audience that gets the full picture than a large audience that only sees the polished version.
How to follow along
The development happens in public, mostly on the Skyline Discord. That’s where we post screenshots of in-progress work, share calibration notes, ask the community for input on specific design questions, and where alpha invitations will eventually come from when the time comes. If any of this sounds like something you’d want to be early on, the Discord is the right place.
We’re also on X (Twitter) for shorter updates and screenshots, useful if you’d rather follow without joining a server.
Devlogs land here, on this site, roughly every two weeks. We’ll cross-post each one in the Discord and on X, but if you want them in your reader, the RSS feed is at /blog/rss.xml and you can subscribe there.
A short note on what we want from you
If you’re someone who’s been waiting for a serious single-player airline simulator, if you played AirlineSim and burned out, if you’ve spent more time than you’d admit looking at Wikipedia route maps, if you read T-100 data for fun, we want to hear from you. We’re going to be asking the community a lot of questions over the next year, both about what to build and about what we’ve gotten wrong. The community we want isn’t passive. It’s the kind that pushes back on bad decisions and helps us think clearly.
If you have aviation industry experience in operations, planning, revenue management, fleet, anything, we especially want to hear from you. We’re going to be running an engine-level alpha with domain advisors in the coming months, and that’s where the simulation’s accuracy gets stress-tested against the people who actually know what airlines look like from the inside.
If you’re a developer, designer, or simulation enthusiast who wants to understand how this thing is built, we’ll be opening up the engine architecture and design decisions through this devlog over time. We’re not open-sourcing the game itself, but we’re going to be transparent about how the pieces fit together.
And if you’re just someone who wants to lurk and see whether this turns into something real, that’s fine too. We’ll be here. The work continues.
Thanks for reading. There’s a lot more to say, and we’ll get to it.
the skyline team