About us

Making games is the fun part of games. We're building the place where making them is the game.

Updated 2026-11-19

TLDR: AI Disclaimer

Every game on Okay What If is made with AI: an LLM, Ohmo, writes its code, and AI makes its simple 3D models and sounds. But only a person can direct the AI, and nothing goes into a game until a person in its project approves it as exactly what they wanted, so a game grows one approved step at a time. It isn't a machine that makes a game from one prompt; it's a tool for building one feature by feature, each change patched in even while the game is being played. Every picture is drawn by hand, by a person, and nothing comes in from outside: the ideas, the choices and the drawings are people's, and that's an intentional design, not a limit of what the AI can do.

TLDR: What's a game on Okay What If?

A game here is made to be hopped into: played with friends for a few minutes or an hour, enjoyed, and left, with no legacy or high score to keep up. Most are multiplayer, and one can be played alone with a single seat; none is an open world with 200 hours of story. Most videos are 10 to 20 minutes long for a reason, for those making them as much as for those watching, and a game you can pick up as easily reaches far more people, which is much of why phone games are played by more people than PC games. We're not saying what games should be, only what they are here. A playground only some can enter isn't much of one, so every game runs in the browser, hosted for you, and any decent laptop, a work laptop included, plays it. Browser games aren't new: what they always lacked was enough of them, and an easy way to make them.

TLDR: Why a feed, not a grid of games?

Home is a feed, one game at a time, not a grid of thumbnails to choose from. People look at games one by one anyway, and scrolling past one takes no effort. A thumbnail can't say what a game is, and with new games made every day, a page of a hundred pictures helps no one find the good ones. A video is easy to choose by its picture, because watching costs little; playing asks more of you, so picking from twenty pictures leaves you less happy with what you picked. Instead, we ask for thirty seconds: a game's intro, its cover, or, best of all, someone playing it live, the plain truth of what it is. Games aren't videos, and we don't treat choosing one as if it were.

Contents

  1. Contact us
  2. More about us
  3. Common questions
  4. What we value
  5. Why we exist
  6. Who it's for
  7. What we believe about AI
  8. How Ohmo works with you
  9. How we make money
  10. What we do with what you make
  11. How the company works
  12. The future
  13. Why it's built this way
  14. These pages
  15. The only question

Contact us

Write to the address for what it's about:

To ask us for a feature, report a bug or ask a question, write a post tagging Okay What If. Everyone can read, reply to and like those posts.

More about us

Common questions

What people ask most, each linked to the place that answers it.

Your games

Making games with Ohmo

Money

Privacy

Safety

These pages, and how we work

What we value

1. Seek the truth

Everything else here follows from this. We do what we do because we think it's the closest to the truth about what a lasting game platform should be: not out of ego, and not to follow a moral code, but because we think it's true. Everything on this page, and every decision behind how Okay What If works, is a conclusion we reached that way, and when we find we were wrong, we change it.

To us, engineering is truth seeking: we say what we think is true, test it, find where it's wrong, fix it, and go again. None of that works if we can't be wrong. Claim nothing and nothing can be shown false, so nothing gets better. That's why we write down what each feature must do as statements a test can prove false. Confidence isn't knowing we're right now. It's knowing we'll get there, by finding out where we're wrong.

So seeking the truth isn't about being right. Showing off what we got right is marketing. Finding what's wrong and fixing it is problem solving, and it starts with the problem.

Be wrong or be right, but never be neither.

Kyle, founder

2. Anyone can make a game

Most people don't spend their days at a computer, and they don't need to: their lives happen away from one. But almost anyone can open YouTube and find a funny cat video. That's the bar we hold ourselves to. Making a game here should take no more knowing about computers than that.

This isn't about anyone lacking skill. Skills are particular, and nobody has all of them. We, the people who build Okay What If, spend most of our waking hours at computers, and even we don't know everything ours can do. Expecting everyone else to learn computers before they can make a game would just be rude.

Give me a hammer and the wood for a birdhouse, and I'll be in an emergency room two hours later. I sit at a computer a hundred hours a week, so I've no business handing people niche tools and calling it a skill issue when they can't make a game.

Kyle, founder

3. Don't be evil

Google once made it a motto, and we're borrowing it, not as a moral rule but as a business one. A platform's most valuable resource is the trust of the people on it: they make what's there, and they can always leave. A company that keeps that trust makes things that last for generations. One that spends it makes money once. That reasoning only holds while people can leave: if we're ever successful enough that leaving gets hard, it stops holding on its own, and we'll need a new rule to replace it then.

So our compass isn't what we could get away with. It's whether a choice fits what the people here value, which is what What we believe about AI is about, and we've put what we owe you in writing, in our terms of service. If keeping to that ever costs us money, we'll pay it, because doing otherwise would cost us more.

If it can scale a multi-trillion-dollar company, it can surely scale us too.

Kyle, founder

Why we exist

I've lived with games for a long time. League of Legends, Minecraft, Stardew Valley, RimWorld, Overwatch, and more than I can list. I've been making games for almost as long, and I have never finished one.

What those unfinished games taught me, and what every developer I know will tell you, is that a game doesn't have to be good, or even finished, for making it to be fun. So I decided that would be the main loop of my game: making games, or at least the fun part of it.

Kyle, founder

"Okay, hear me out."

That sentence, said on a call with your friends, should be all it takes to make a game. That is the whole company.

Capability, and experience

Every product has two sides: what it can do, and what it's like to use. What AI can do isn't our advantage. Everyone can buy the same models, and they get better for everyone at once. What nobody seems to have figured out is the other side: what it's like to make something with one.

The trouble with a token sink

Every company building on AI wants a token sink: something people keep spending tokens on. The trouble is where the value comes from.

Prompting usually isn't fun. Developers everywhere talk about the death of programming, and about how draining it is to prompt, read what comes back, then prompt, and prompt, and prompt again. It's a bad experience.

When making is a chore, the money has to come from something else: people consuming what was made. For games, that ties every token spent to one question, whether the game is fun to play. That's a big risk:

Even if an AI can drive a game engine through MCP and make a good game, that only serves the people who play it. It misses an entire market of value: the making itself.

Fun to make, fun to play

Not all games that are fun to make are fun to play, but all fun-to-play games are fun to make.

This isn't a heartwarming philosophy. It's where the value is. If making a game is itself fun, people don't pay to make money, or to end up with a product. They pay to entertain themselves, the way people pay for any game, and that's a different business.

And people who like prompting keep prompting whether or not what they made turns out to be fun, because that was never the goal. The journey is.

The soul is in the making

Everyone has seen it by now: games made with AI that all look alike and have no soul. We don't think the soul is missing because an AI helped. We think it's missing because nobody had fun making them.

To us, the soul of a game is everything that happens while you make it:

It was never about AI, and it isn't about being effortless. Nothing is effortless, and you pay for every token. A good game its author never loved always turns out worse than a bad game its author did.

A game is a story its author writes. Every experience in it is intended and shaped by someone. Slop has no soul because it was written for the player, not by an author who cared about it. That isn't to say games shouldn't be made for money. It's to say what ours is for: taking you through the journey of making games.

Making games is play

Ask a game developer why they chose this life. It wasn't for the money: most of them could have made more doing something else. They chose it to share the experience of play, and above all the kind only games give. Almost every game developer was inspired by a game someone else made. Making one is itself a kind of play: you have an idea, you make it real, you watch someone laugh at it. Developers have always wanted everyone to feel that. That is our directive.

We say everyone should make games the way you'd say everyone should try singing, or cooking for friends. Not to make money, and not as a career. Because it's fun and rewarding.

People have always had ideas. Social media didn't make anyone more creative. It made acting on a thought nearly free: you think something, and a minute later it's posted. Everyone has an "okay, hear me out" idea. Most people have just found making it too expensive: in time, in money, in skill, in patience.

Nobody is paid to post on Instagram or X, and billions of people do it anyway, because making something and sharing it with people you know is one of the oldest joys there is. We're betting that making games belongs on that list, and that the demand is there, unrealized, among people who would love making games and have never tried, because it has always cost too much.

If we're wrong about that, Okay What If is doomed, and so am I. That's the bet.

Kyle, founder

So our job is simple to say and hard to do: make prompting fun, and make it frictionless. If you think it, you should be able to make it tonight. Take the expense away, and people act on what they think.

It's a game

Under the hood, Okay What If is a game engine. We don't think of it as one, and we hope you won't either. It's a game, and the game is making games.

Think of a board game. Someone brings the box and sets it up. Friends sit down and play. When the night is over, the box goes back on the shelf. Nobody hosts game night to make money. They host it to play.

Okay What If works the same way. Someone opens a project and starts the game. Everyone is in the same running world, on a call, talking over each other. Someone says "okay, hear me out: what if the floor was lava?", types it, and a minute later the floor is lava. Everyone is still standing in the game when it happens. You drag a slider until the lava is exactly as slow as it is funny. Someone else says "okay, but think about it…", and the next idea goes in. When you're done, you turn it off.

The game you pictured at the start is not the prize. The journey is.

Asking Ohmo is giving orders, not waiting on them. You're the one with the plans, scheming and watching them carried out, never a developer waiting for a prompt to finish so you can push the next. And every version is progress you can feel: the goal is the next step, not "make a game".

We don't sell games, and neither do our users. We sell the experience of making them together.

Who it's for

Okay What If isn't designed for game developers. Unity, Unreal and Godot are excellent professional tools, and their business depends on studios shipping games that make money. That shapes everything about them: you install them, learn a programming language, and spend weeks before anything is playable. The finish line is a store page. They're tools for work. Okay What If is built for play, and that changes the machine underneath.

If you make games already, you're welcome here, but we'll be honest: if you write code, one of those engines, connected to an AI assistant through MCP, will give you far more control than we ever will.

We're built for everyone who wants to make games without any of that: nothing to install, since your game runs in the cloud and players join from a browser; no engine to learn, no code to write, no AI tools to connect. If you can describe it to a friend, you can make it here.

Okay What If is made for friend groups and the Discord server you already have, not for strangers.

And most people here will come to play. There will always be more players than makers, so playing, and talking about what's played, get the most care, and playing should take no effort at all: people expect a lot of it now. So outside the Editor, the site works the way the social apps people already know do, YouTube, X and Instagram among them, and every button does what people expect of it.

What we believe about AI

To us, AI is neither good nor evil: it's a tool. We ask only whether it adds anything of value, and for what you do here, value means fun, since play is what we're for. Where it gets you to the fun part sooner, the drawing and the scheming, we use it. Where it takes the fun away, as watching it build your whole game while the bill climbs would, we don't.

Life is full of decisions. Some are chores and some are interesting. People pay to skip the chores and to keep the interesting ones, and we think that second kind is where AI will find most of the people who pay for it.

It isn't the AI that's misaligned

People worry that AI isn't aligned with human values. We think the models are closer to our values than they're given credit for. What's out of line is the business around them: the people building AI products have misread what people value. They've misread which decisions people want made for them, and which they want to make themselves.

It's the same with any tool. A hammer can build a house or hurt someone, and nobody asks whether the hammer was a good invention, or says everyone who uses one is a bad person: the good and the harm are the work of whoever holds it. We see AI the same way, and look at each use in good faith: what it's meant to do, and what it brings about.

Interesting decisions

Everything in life is made of decisions, even the small things. Some have a right answer: how to file your taxes, how to wash a shirt, how to word a work email. There's little to choose between, so people happily pay to have them done. Others have no right answer: what your game is about, what happens when someone touches the lava, who you play it with. Those are the interesting ones, and they're what people do for fun. Game designers have long said that a game is a series of interesting choices.

The same task can hold both kinds. Making a game means thousands of decisions in its code, and most people find those tedious. Programmers enjoy them, but programmers are few. What the game is, though, is interesting to everyone.

Good uncertainty, and bad

We think much of what makes something fun is uncertainty, and the moment it's resolved: you choose, you find out how it went, and you choose again. That loop is where the rush of dopamine is.

There are two kinds of uncertainty. The good kind is not knowing how your choice will turn out. The bad kind is not knowing whether the tool will do what you asked, which feels like a slot machine: there's nothing to choose, only something to wait for. A prompt that goes straight to a finished result has neither choice nor discovery, so nothing is felt. The products people love bring the next good decision closer: scrolling asks you one every few seconds, watch this or swipe past it?

Nobody wants to skip their life

Ask almost anyone whether they'd rather skip to the end of their life and know how it turned out, or live through it. They'll choose to live it. The value was never the ending, dying at 99 with a loving family. It's the living: the ups and downs, and the memories, which only come from being there.

People pay for experiences: a concert, a trip, a game. The pitch behind much of AI, and behind AGI most of all, is a tool that does the living for you. That asks people to pay to have their most interesting decisions taken from them.

That's just not how human value works.

Kyle, founder

Good results come from many tries

Picture two tools. One is miserable to use, but what it makes is good. The other is fun to use, but most of what it makes is bad. Industry asks for the first. People's wallets open for the second.

The second wins because the first makes a promise nobody can keep. Most of what anyone makes doesn't work out: most games, most songs, most companies. A good result is the rare one among many tries, and the tries are what lead to it. So what matters is making each try quick, easy and enjoyable, since more tries means more of the ones that work.

YouTube shows it best. Most of what's uploaded to it isn't good, and never was, long before AI. Yet it's one of the most successful platforms there has ever been, because making a video is easy enough that millions are made, and the good ones find their audience. People pay for YouTube Premium because there are videos worth watching, not for a chance at making them. Nobody pays to take on the risk of making something; they pay for what came of it.

Platforms don't make what's on them

Most AI companies are aiming at a small share of the world: the work that needs precision, where reliable, fast results make money. But the platforms that last, YouTube, X and Steam among them, for the most part don't make what's on them. Their people do: Steam began as Valve's way to deliver its own games, and lasted because others could put theirs on it too. A company that carries the cost of making everything itself is a provider, not a platform.

For most of the world to pay for AI, it has to be something they enjoy, because everyone plays: a grandmother who never touches a spreadsheet still watches YouTube. AI aimed at work will matter, but to a few. Much of the money going into AI is a bet that ordinary people will pay hundreds of dollars a month for it, and we think that only happens if it's fun.

When the chores are gone

Some AI already works this way, and we think it's right: assistants that take the tedious decisions off your hands, like which of a hundred chargers to order, when to set an alarm, or what's still left on a list. If they succeed, we're better off. Every chore taken away is free time, and people spend free time looking for something to decide: a game, a sport, a puzzle.

That's what people are, in a way: decision makers. Take decisions away and they go looking for more, and when they run out, they invent problems just to solve them. Ask a mathematician. So we don't think AI will ever replace people, however many decisions it takes on: people will always find new ones to make.

That's where we stand. Okay What If is where the interesting decisions go: in games, and in making them.

What we do about it

Ohmo makes the decisions most people find tedious, its code, and leaves you the interesting ones: what your game is, and what happens next. It asks before it builds anything bigger than one step, and what your game becomes is your call (How Ohmo works with you). Our job is to get you to your next interesting decision as quickly as we can, because that's where the fun is (Why we exist).

We think this is where most of the public's dislike of AI comes from too. Few people hate the technology itself: nearly everything anyone uses now leans on AI somewhere, so the real question was always how it's used. That doesn't mean they like everything it brings. What they dislike is what's being sold to them, which doesn't fit what they value. People want to live.

Where the fun draws the line

Fun also decides what AI makes here. Drawing is part of the fun of making a game, and everyone can draw: nearly every school teaches some, and people drew on cave walls. So you draw your game's pictures, and AI doesn't. 3D modelling is hard, and learning it is too expensive for most people, so AI makes 3D models, animations included, and no 3D game is held up. But only simple, low-poly ones, which you can colour as you like: a flawless model made for you would leave you nothing to make your own. Low poly also keeps games running on almost any computer, and leaves room above it for the 3D artists we want here one day.

For sound, AI writes a program that makes simple sound effects and loops from a description, and that's as far as it goes: no voice lines, no dubbing. Your own recordings and music come later: we can't yet check what comes in for safety and copyright the way a video site does, or make sound a full feature (How we keep to it).

None of this is a promise. If people would rather AI did more, we'll move the line, and say so first. But people optimize the fun out of things when they can, so where it falls stays a judgement call, made as we go, to keep making games fun.

Not ours to decide

If we decided AI was wrong, say for art, Okay What If would be deciding what's moral for everyone on it, whatever their faith, beliefs and values. That isn't ours to decide. AI stays out of your drawings because it would take the fun away, not because it's wrong. Like everything here, what we think of AI follows from what we think is true, not from a moral code (Seek the truth).

Here to stay

Whatever becomes of Okay What If, we think AI is a real advance, and it'll keep getting better and more widespread. It has permanently changed how software works, and much else besides, and there's no going back to how things were: not because we won't, but because we can't.

Humans are really good optimization monsters.

Kyle, founder

We run on data centers too

We think computing, and the energy behind it, will be one of the most important commodities of the years ahead. For us it already works like a gold reserve: it's what most of our money buys, priced in the currency of whoever sells it. The AI is bought in US dollars from an American company, and the servers our games run on, in Canada, are priced in US dollars and paid in Canadian ones. Our prices are in US dollars too (Plans and credits), so what we earn and what we spend move together. We're betting we'll keep buying it from companies in Canada and the US: if we ever had to buy it in another currency, our finances would change from the ground up.

So we want the data centers we depend on to be in Canada and the US. That doesn't mean forcing them into anyone's backyard: governments, businesses and the people who live near data centers have to find a way that works for everyone. That's a hard problem, and we don't have the answer.

Okay What If runs on data centers, so we can't claim innocence. The criticism that may come with them, from the strain they put on power grids to what they do to the places around them, falls on anyone who uses software and the internet, us first, and we fully accept it. Maybe it's less a criticism of a select few than of all of us.

Where we could be wrong

If AI one day outdoes people at everything, and costs less than hiring one, then we're wrong, and far more than this changes for everyone. Whether that day comes, we honestly don't know enough to say: that's a question for the people building the models at the big AI companies, not for us. What we've written here is our opinion, about what people want.

How Ohmo works with you

Ohmo, for Okay, hear me out, is Okay What If's AI assistant, the one you make games with: you describe what you want, and it builds it, checks it and hands you a new version to play. In technical terms, it's an agent working through the harness the platform gives it. Here is what it does, and how it checks its own work. What it will always build for you, and what it won't, are in our terms of service.

It talks back, in three ways

Nothing is built until you sign

You write Ohmo a prompt, and it answers you, or proposes the tasks your change takes, a feature each: what it understood, how it splits the work, what it's unsure of, and about what each costs. Nothing is built until you approve a task. Approved tasks wait in your project's queue and are built one at a time, while you do anything else: each comes back to you, and the next begins only once you've signed it off as what you wanted. Deciding and imagining is the fun part, and we don't want to take it from you.

How it checks its work

An AI that writes code is easy to find. An AI whose work you can trust is not. Before a change reaches you, Ohmo must:

If something fails, it fixes the problem before you ever see it. Ohmo also looks at pictures of its tests, drawn by the same renderer you play with. A check guards only what someone thought to check; a picture catches what nobody did. You are the best judge there is of whether a game is what you wanted, so your approval outranks Ohmo's: when a change redraws a picture you once approved, it comes back to you to approve again. A test that passed yesterday still means something today, because every run of a game is the same. How we test covers the checks beyond Ohmo's own.

Every behaviour comes with a test you can play

When Ohmo builds something (a jump, a coin, a door), it writes a test of it, kept once you say it's what you had in mind: the moment itself, set up in your game. You can start that moment in your own editor window and play it yourself. You don't have to take the AI's word for it: you can press play on the proof.

The AI underneath

Okay What If was built around Claude, Anthropic's AI. It isn't tied to it: models from other providers can work too.

We hand-pick the models Ohmo runs on, and decide whether to offer you a choice between them. That isn't to make you spend more tokens, or to make as much money as we can: it's to make sure every model it runs on is capable enough to clear the minimum bar of what making games here should feel like. As better models arrive, we'll choose and provide them in good faith, though that isn't a legally binding promise.

How we make money

We pass on what the AI costs us at exactly the rates its providers charge, and add nothing to it. Half of what your plan costs comes back to you as credits for it; the other half pays for everything else that keeps Okay What If running, the machines your games run on included. Host, the least plan, gives $10 of credits and pays chiefly for the two pods it adds to the public games' pool.

Pass-through, and half of what you pay

We're a middleman and a router. We buy AI from model providers, route it to your game, and pass it on to you at exactly what it costs us. The machine time your games run on, bought from a cloud provider, we don't pass on. Okay What If makes its money once, from your plan: the half we keep pays for taking your payment and for everything we cover, your games' servers included. What a plan gives, and what you pay for, are in our terms of service.

What we promise to do with what we keep is written in our terms of service.

Who pays for what's free

Everything here is paid for by someone. Playing is free, and so are the servers public games run on, because subscriptions pay for both: public games run on a pool of servers, four of our own and two more for every subscriber. A subscriber never waits for a server to play a public game: when they run short, the smallest lobby with no subscriber in it makes way.

That's also why there are no ads here, and why nothing anyone does is sold or shared for advertising: today, subscribers pay that price for everyone. It's an arrangement, not a principle. If people would rather have a cheaper or free way in, paid for by ads or something else, and it serves them better, we'll offer it, and the privacy policy will say so before anything changes.

And a plan can't simply get cheaper. What a task costs is what the AI's providers charge for it, which moves with every model; building it into a plan's price would mean betting on it or paying for it ourselves. So a plan's price stays about where it is, and its credits pass the AI's cost on as it is.

How we see you, from where we sit

As a business, you are volume. Every change you ask for passes through us. We aren't trying to get a lot of money from you once. We make money when a lot of people have fun often. So everything we build pushes the same way: cheaper changes, faster changes, fewer steps between an idea and seeing it. When a change gets cheaper, you make more of them, and that is exactly what we want. How we hope to bring the price of the AI down is in What we do with what you make.

Scale or die

Keeping less of what you pay, and still making a real profit, takes an enormous number of people making games. That isn't a slogan; it's the arithmetic. Either making games becomes a normal thing people do on a weekend with their friends and family, or Okay What If never becomes what we're building it to be.

Why take a bet like that? Honestly, because we love money. We love it so much that we'd rather build the version where we make an absolute bank than a safe little business, and we know the other side of that bet is going bankrupt.

And that's okay: I know a secret tech where if you're self-employed, you'll never be unemployed 🀨😭😭😭😭

Kyle, founder

We're telling you all this because it's true, and because it means our interests and yours point the same way: we only win if you keep coming back because it's fun. Where our own money comes from today is in How the company works.

What we do with what you make

What you make here, your prompts and the work Ohmo did on them, your games, your posts and your snapshots, may be used to make AI, and Okay What If, better and cheaper. We may train AI models on it, ours or another company's; we may give or sell it to the companies that make AI models; and we use it to study and improve Okay What If. We never publish your private projects, and never claim your work as our own.

Why. Most of what a night of making games costs is the AI, priced by the companies that make it. The record of millions of people making games, how they really ask for things and how each request turned out, is worth something to those companies, and could train a model that does Ohmo's work for less. That's how we hope to bring the price down.

You can always opt out, in your profile's Settings tab, and from then on nothing of yours is used for any of it. Before anything of yours is first given or sold to anyone, we tell you at least 30 days ahead, so you can opt out before it goes. What's used, what never is, and who it goes to are in our privacy policy.

Scripts are the exception. Part of what Ohmo does on your tasks improves the scripts Okay What If keeps for every game, from the logic it writes, never your words or anything else that's yours alone. You pay for some of it in your tasks, and it's what lets every task cost less later. There's no opting out of it, since every game is built on those scripts.

Don't expect a halo

We'd rather you didn't expect more of us than this. No ads, nothing sold for advertising, what we do with what you make: none of it is a moral stand we identify with. It's how Okay What If works best today, and it changes when what the platform needs does. Our terms of service and privacy policy change with it, and tell you before a change takes effect. We're not a company claiming to be good: we'll do whatever works best, and say so here. If you don't like what that is, opt out, or leave. What doesn't change is Don't be evil: whatever we do with what you make, you're told first, and your opt-out is kept to the letter. That's what keeps it fair.

How the company works

How we're organized and funded, and how help reaches you. You should be able to see how the company behind your games is run.

Today, the company is one person: Kyle, the founder, with Claude writing most of the code. This page describes how it runs from launch, as it grows. When something here changes, this page changes with it.

How we're organized

Okay What If is made by teams. Each team is responsible for one part of Okay What If, and each has one or two leaders who answer directly to the founder. There is no separate HR department: a team's leaders decide who joins their team, and the founder approves the funding for it.

The teams follow the parts of the system:

Security looks after the code as a whole. Architecture shows where each team's work sits in the system.

How we're funded

We have no partners and no funding. Okay What If is bootstrapped, with the founder's own money and $1,000 in AWS Activate credits. That's everything. The future says what we're looking for.

Everything goes through the app

The app, the site you're on, is where you make and play your games. It is also where most of what you ask of us happens: feedback, and anything to do with the rules we follow. Help with an account, a payment or a refund, a request about your data, and business with us go by email, to the addresses on About us.

When you need help

Our customer support staff answer what you write to us from our own employee terminal, where each email is a case until it's resolved. On the site, their accounts can do no more than a player's, and they have no access to our source code.

When support can't resolve a case, they escalate it to the team responsible for that part of Okay What If, and that team takes it from there.

The future

What 1.0 is, what we want after it, and what we haven't decided yet. We won't release Okay What If until it's ready.

What 1.0 is

How we launch

We launch in Canada and the US together, with everything the US needs in place: its laws kept. From launch, people in both countries sign up, play, post, make games together and buy a plan, which Stripe sells for us in both, handling its sales tax.

Why buying opens in both at once. The credits a plan gives are yours to spend as soon as you buy it. We're paid for the plan about a week later, and billed for the AI you spend them on only at the end of the month, so we're paid before we pay for what you spend, and both countries can open at once.

Growing past one person

The first thing we want is people:

Funding

To get where we want to go, we're looking for funding: investors, or acquisition by a company that can run this for millions of people. The natural partners are cloud providers like AWS, and AI model providers who want their models to be the ones running behind millions of game nights. If that happens, you'll read about it here. How the company works says how we're funded today.

Further out

Waiting on a decision

Features we've thought through, and chosen not to decide yet:

Why it's built this way

Why Okay What If is built the way it is: the engine every game runs on, the checks every change passes, and how the whole system is arranged.

The engine

Every game here runs on an engine of our own. This is why we built it.

A small set of pieces. Games are made from a small set of pieces that repeat in every game ever made: things that move, things that touch, things that change, timers, controls. Today's AI is already good enough to put those pieces together. What was missing was the environment around it. That's what we built: a game engine of our own, multiplayer from the start, and the checks around the AI (How we test).

Every run is the same. We made the engine deterministic, chance included, so that a test means something: a game that passed its test plays the same way when you play it. That's what lets us promise that the engine never breaks your game.

Its physics is our own for the same reason: a physics library made for speed doesn't promise the same result on every machine, and every test rests on that.

Every game is multiplayer from the start. A game was always meant to be played with other people, so the engine was built for it from the start: a game made alone is played together without being rebuilt.

How we test

Every change passes through checks before it counts as done: Ohmo's own, for each change to a game (How it checks its work), and ours, for each change to Okay What If itself, for the reason Seek the truth gives. Here is how we keep ours honest.

Tested as a person would. A test for each feature drives the site the way a person would, in a real browser, with the graphics card in use so games really render, because a page can look right while what's behind it is broken. It acts like a person, checks what a person would notice rather than what usually comes with it, and looks behind the page too, at the site's own routes, its database and its running servers, so what the screen shows is also what's true.

Keeping tests honest. A new check is seen failing once before it's trusted: a check that has never failed may not be able to.

Every game's tests, on every engine. Every test of every game people make is run on each new engine, and plays the same way every time. When an engine fix changes what a test recorded, because the test recorded the old fault, the engine stays fixed: the test is marked wrong and retired, counted on Statistics, and the version holding it is marked for its owner to fix with Ohmo.

Architecture

Data belongs to its handler. The site and the system that runs its games each have one keeper for their data, with a database, storage and credentials of its own, so a fault on one side can't reach the other's data; and since each keeper is the only thing that knows where its data is, deleting a person's data happens in one place.

A directory is a job. The repository's top-level directories are job titles, and each holds exactly what that team owns, so handing someone a directory is handing them the job. What each team is responsible for is in How the company works.

Arranged to be read. Most of our code is written by AI, and all of it is read far more often than it's written, by people and by models alike. So it's held to one measure: what it costs a reader to find, understand and change it. A reader doesn't take in the whole repository; they look at a folder, choose the branch most likely to hold what they need, and go on. So the tree is arranged for the likeliest choice at every step to be the right one, and the same wrong turn taken again means the arrangement is wrong, not the reader. A reader who finds their way in the end proves nothing: their persistence hides the fault they worked around.

That's why one kind of shortcut is allowed and another isn't. A slower way of doing something, kept inside the file that does it, costs only that file's readers, and can be fixed whenever it matters. A file in the wrong place, or a name that doesn't say what it holds, costs every reader after it, so it isn't allowed.

These pages

Written by AI. This page, our privacy policy and our terms of service are written by AI: Claude, the same AI that writes most of our code. The decisions in them are ours, not the AI's. Claude puts them into words, and keeps the words true as the product changes.

Why we say so

A company that makes games with AI shouldn't pretend its own words came from somewhere else. What matters is that what's written here is true, and that someone answers for it. Every decision on these pages is intended, and ours, not the AI's.

Why we write it all down

This isn't a brochure. It's the plain truth of who we are, why we exist, what you pay for and what we promise, written down in plain words so you can hold us to it. Some of it will read as obvious. We write the obvious down anyway, because a rule nobody wrote down does not exist.

We also tell you what most companies keep to themselves: how we see you as a business, what you pay for, what we take, and what happens to us if this doesn't work. After all, limits and values shape a product, and a healthy relationship comes with clear boundaries.

These pages are public because the relationship we want with you is one you can see, check, and hold us to.

How it's kept

Each fact is written once, on the page it concerns: this one, our privacy policy or our terms of service; every other place that needs it links there. So no two pages can tell you different things.

A page changes when the product or a policy does, and shows the date it last changed at the top. When a promise changes, its page changes first.

What isn't here

We're open about how we work, but we're not open source. These pages stop short of what would put people or the product at risk: our source code, the details of how we keep things secure, and the engine's book, all Ohmo knows of the engine.

The only question

The engine builds the worlds. Your friends make them worth building.

So here I am, asking you the only question Okay What If exists to ask: "Okay, hear me out."

Kyle, founder