Coinpath Template: A Budget App You Can Rewrite by Chat

Coinpath Template: A Budget App You Can Rewrite by Chat

Coinpath is one of the 25 ready-made Zugo templates: a personal budget app with three real screens, a marketing page, a sign-in page, and a dashboard holding transactions, categories and a monthly trend. Cloning it into your project costs no credits, and from there you reshape it with plain sentences in chat.

Below is what actually ships inside the file, which changes it survives without breaking, what you still have to connect yourself, and where the template honestly stops being the right starting point.

What is inside the Coinpath template?

Coinpath sits in the app category rather than the landing category, which means it carries state instead of just copy. The file contains three screens the project navigates between on its own, and that three screen shape is shared by every non game template in the library.

The transaction list has a three way filter: all, in, out. The demo categories are domestic ones (housing, groceries, dining, transport, utilities, health, fun, salary, freelance and other), and there is a working empty state, which matters more than it sounds because the empty state is the first thing a new user meets.

One detail worth keeping: the sidebar labels for trends and savings goals are decoration in the demo. The dashboard is a single screen, not five. If you want those as real pages, say so in an edit and they get built.

Who does Coinpath actually suit?

The obvious fit is personal money tracking: a monthly budget, spending review, saving toward a goal, a subscription audit. Balance on top, movements below, category split beside them works for all four without rearranging anything.

The second fit is a small internal tool. Project spend, a department limit, per diem reconciliation. The screens stay identical and only the labels and the owner of the data change, which is a cheap edit rather than a rebuild.

The third is less obvious. Coinpath works as a frame for any app that has one running number and a list of events feeding it: calories, kilometres, billable hours. The mechanics are the same and only the unit differs. If your case is closer to done or not done, the Continuum template covers habits and streaks instead.

Where it does not fit is bookkeeping with double entry, tax obligations and multi user permissions. That is a different class of product, and stretching this template that far costs more than describing your own from scratch.

How does a template differ from a blank prompt?

The difference is not output quality, it is the starting point. A blank prompt gives you a project reasoned out from your description. A template gives you a project reasoned out in advance, and you get to see the whole thing before spending a single credit.

Two practical consequences follow. First, the structural gamble disappears. In Coinpath it is already settled that the monthly balance sits above the list and the category split comes after it, and you can open that decision and poke at it rather than imagining it.

Second, the cost of a wrong first try. A build from scratch costs 6 credits, and if version one misses, version two costs 6 again. The clone costs nothing, so choosing badly is free: it does not fit, you try another template.

There is a trade in the other direction. If your case sits far from income and spending, adapting is dearer than describing, because you pay to remove someone else's decisions on top of adding your own. Simple rule: if three edits in you are still tearing out the original idea, start fresh.

How do you adapt Coinpath to your own case?

The usual chat rhythm applies: one sentence, one clear goal. Small steps land faster and cheaper than one enormous prompt, and you can see where each credit went.

1. Change the domain first: "rename it to Kilometro, count kilometres instead of money, replace the categories with running, cycling and walking". A single edit touches the marketing page and the dashboard together.
2. Fix the categories and the currency: "keep four categories, rent, food, transport and other, and show amounts in dollars". Categories are the first thing that drifts from a real user's life.
3. Cut what you do not need: "drop the month over month chart and show my five largest expenses instead". Nothing in the template resists being removed.
4. Work on the empty state: "when there are no transactions, show a hint with an example of the first entry". Cheap edit, large effect on first impressions.
5. Publish to your own address on .zugo.run and open it on a phone. Money screens get checked on phones far more often than on laptops.

Review after each edit rather than in one batch at the end. Before anything reaches you, every build is opened in a sandbox first, so a broken version is reported as a failure instead of landing in your project. The general mechanics of chat editing are covered in how to edit after generation.

What does adapting Coinpath cost in credits?

Cloning the template consumes nothing: the page is copied into your project without calling the model. Billing starts at your first edit.

Free gives you 5 credits, which covers the clone and one edit. Pro is $25 a month with 200 credits, roughly 66 edits or 33 builds. Business is $99 a month with 800 credits. For planning: making a template feel like yours usually takes between five and fifteen edits at 3 credits each.

What has to be connected before the budget is real?

The template is an interface with demo data. Entries survive a page reload in the browser, but they are not yet tied to an account, which is the difference between a demo and a product.

Supabase covers both halves: tables for transactions, email sign in, and file storage if you ever want receipts attached. Ask for it as its own edit: "connect Supabase, store transactions in a table, and show only the current user's rows". That last clause is the important part of the sentence.

After that, as needed: Stripe if the app is paid, Resend if a monthly summary goes out by email, Google Analytics if you want to know which screen new users stop at. GitHub export hands you a real repository, so the code is not trapped.

Order matters. Database and sign in first, everything else after. The reverse usually means analytics measuring behaviour on data that is about to vanish. The longer version lives in can AI build a database app.

What does Coinpath not do?

Three honest limits, worth knowing before you start rather than after.

There is no bank connection. Transactions are typed in by hand, and the marketing line about not asking for bank access describes the demo rather than a hidden feature. Importing statements automatically is an integration with an outside provider and gets built separately.

The demo figures are decorative. The name, the month and the amounts exist to show the screen. Until a database is wired in, the app looks like it works but only remembers what one browser remembers.

Serious financial logic arrives through edits, not through one giant prompt. Recurring payments, carrying a balance between months, per category limits with alerts: each is its own conversation. Zugo does not replace a development team on a complex product, and handling other people's money turns into a complex product quickly.

None of that is a reason to skip the template. It is a reason to know where it ends: as a personal tool and an internal tracker it covers the whole job.

Where to start

Open the template gallery, find Coinpath under apps, and run the preview. Because the clone is free, clicking through the screens costs nothing, so do it before writing your first edit. Make that first edit about the domain rather than the colours, or you will pay for both twice.

Then move in short steps: one edit, one look, one decision about the next move. Credit spend stays legible that way, and the project does not wander somewhere you have to drag it back from.

You can take the template and start rewriting it at zugo.dev. Free credits get you to a first version that is recognisably yours, and the published project gets an address you can already show someone.