A challenge is a simple thing: pick a problem worth solving, invite people to build something that solves it, and put the results where everyone can use them. We are looking for organisers anywhere in the world.
How a challenge works Run one where you areEvery entry is one HTML file that opens in a browser. That constraint is not a technical preference. It is what makes the competition fair.
Take away the file and you are no longer judging the idea. A conventional app needs hosting, a database, a build pipeline, a deployment process and someone who knows how to operate all of it. A team with those things wins before anyone looks at what they built. A sixteen-year-old with a good idea and no infrastructure does not get to compete at all.
One file removes every one of those advantages. There is nothing to deploy and nothing to pay for. The person who understands the problem best has as good a chance as the person who has done this professionally for a decade — and in our experience they often do better, because knowing what the tool should do is the harder half.
It also means anyone can judge it. A file opens on the judge's own machine, in seconds, with no account, no demo environment and no risk that the server was down on the afternoon they looked.
Entrants describe what they want to an AI assistant and it writes the software. The skill being tested is understanding the problem, which is the skill that was always scarce.
No hosting, no licence, no cloud account, no card. The barrier to entering is a browser, which every entrant already has.
These apps run in the browser and transmit nothing. Someone's household budget or client list never leaves their own machine — which matters when the subject is money.
No dependencies to rot and no service to withdraw it. A winning entry keeps working long after the challenge is over, which is the difference between a competition and a contribution.
The whole program is in one file a judge can open and read. Nothing is hidden — which is the only real basis for trusting a financial tool somebody else wrote.
A tool built for one place works anywhere with a browser. The winner of a challenge in one country is immediately useful in another, which is why we want to compare them.
Households budget everywhere. Small firms keep books everywhere. But the rules, the products, the tax treatment and the habits differ enormously — and so do the solutions people reach for.
That variation is the interesting part. A budgeting tool built for a country where most people are paid weekly in cash looks nothing like one built where everything is a direct debit. Neither is wrong. Each is right about the place it came from, and each has something to teach the other.
So we are not trying to run one global competition and crown one winner. We would rather see many local challenges, each answering a question that matters where it is held, with the results published openly so they can be compared, borrowed from and improved on.
If you organise one, the entries stay yours and stay attributed to the people who built them. What we offer is somewhere to publish them, a format that makes them comparable, and an audience beyond your own region.
This is not a detail in the rules. It is the thing that makes a challenge worth running rather than just an event that happened.
A tool that solves a real problem should reach the people who have it. If the winning entry of a challenge sits in an archive, the challenge produced a result and no benefit. Published openly, it is used by people who never heard of the competition, in places the organiser has never been.
And comparing requires reading. We want challenges in different countries so the answers can be set beside each other — but you cannot learn how someone solved a problem from a screenshot. You need the file, and you need permission to take what is good from it. An open licence is what makes that legal rather than merely possible.
The licence requires your copyright notice to travel with every copy. Anyone who uses your work has to keep your name on it. That is a stronger guarantee of credit than most competitions offer.
You are granting permission, not handing over ownership. Neither the organiser nor Globefin acquires your work. You can keep developing it, sell services around it, or use it however you like.
An entry that cannot be opened and inspected cannot be judged fairly. The whole program is one readable file, which is also how a judge checks the arithmetic is right rather than taking a demo on trust.
Someone in another country picks up your entry, adapts it to their rules and publishes the better version — with your name still on it. That is the outcome we are actually hoping for.
Say this clearly when you announce a challenge. Entrants should know before they start, not after they have built something. Anyone who wants to keep their work closed should not enter — that is a reasonable choice, and it is far better made at the beginning.
These are starting points, not a menu. The best challenge is one where you already know the problem is real, because you have watched people struggle with it.
A tool that helps an ordinary household see where the money goes, plan for something, or get out of trouble.
A budget that works with irregular income. A debt-payoff planner. A first-job starter kit. Something for a household that has never had a spare month.
A shop, a workshop or a market stall keeping its own books without an accountant on retainer.
Double-entry books for a one-person business. Invoicing and chasing. Cash-flow forecasting for a firm with seasonal trade.
Keeping track of customers, members or cases without paying per seat per month for software built for far larger organisations.
A client tracker for a two-person practice. A membership list for a community organisation. Follow-ups for a sole trader.
Tools that make a professional technique usable by someone who was never taught it formally.
A discounted cash flow anyone can follow. A lease-versus-buy comparison. A model that shows its own assumptions.
Not a calculator but an explainer — something that makes a concept click by letting people move the parts.
Compound interest you can feel. Why diversification works. What inflation does to a fixed income over thirty years.
The most interesting challenges are the ones we would not have thought of, because they come from a problem specific to somewhere we have never worked.
Remittances. Informal lending circles. Smallholder farm accounts. Currency instability. Tell us what it is.
The rest is detail. We can help with all of it, and we have run one ourselves, so the advice comes from having made the mistakes.
One sentence describing what a good entry would do for a real person. Narrow beats broad. "A budgeting tool" produces a hundred vague entries; "a budgeting tool for a household paid weekly in cash" produces entries you can actually compare.
Students at one institution, anyone in a region, a profession, or genuinely anyone. Wider is not always better — a challenge open to one college produces entrants who know each other, which makes the judging and the sharing afterwards far more useful.
Say the criteria before anyone builds anything. Ours were: does it work, does it help a real person, and can a stranger understand it. Publish the criteria with the challenge, so entrants build toward the thing you actually want.
Announce the open licence at the same time. Entrants should know their work will be published before they start building it.
Long enough that people can think, short enough that they start. Somewhere between a month and a season tends to work. Anything longer and entrants forget; anything shorter and only the people who were already free can take part.
Tell us who you are, where you are, and what problem you think is worth solving. We will come back to you with what we learned running ours, and help you set yours up.
Express interest See what people buildNot ready to organise one? Browse the apps people have already made — or build something yourself and see how it feels first.
No, and that is the point. Entrants describe what they want to an AI assistant and it writes the software. What is being tested is whether they understand the problem well enough to say what the tool should do.
In practice the strongest entries often come from people who have never written a line of code but have lived with the problem for years. How that works →
Because it is the only constraint that makes the contest fair. Anything more elaborate rewards the entrant who already has infrastructure, and quietly excludes the one who does not.
It also means every entry can be judged, downloaded, kept and used — by anyone, on any machine, without your organisation having to host anything.
The people who made them. Entering grants permission for others to use and build on the work; it does not transfer ownership. An organiser does not acquire an entry by running a challenge, and neither do we.
Entries are published under the MIT Licence, which requires the author's copyright notice to stay with every copy. Why that is a condition of entering →
Nothing on our side. There is no fee to publish entries here, no hosting to pay for, and no charge to entrants.
Whatever you choose to do beyond that — recognition, an event, prizes — is yours to decide and yours to arrange.
Our focus is finance, because that is what we know well enough to be useful about. But a challenge about something else is welcome, and we would rather see a good challenge on a subject we know less about than a weak one on a subject we know.
It happens, and it is almost always the question rather than the audience. A challenge that is too broad, or too obviously aimed at people who already build software, gets very few entries.
Tell us the question you are considering before you announce it. That one conversation is the most useful thing we can offer.