Can AI build a bill-splitting app like Splitwise?

Build a personal bill-splitting web app to replace Splitwise Pro for my own use. I enter all expenses myself and share summaries.

Tech rules:
- Plain HTML, CSS, and vanilla JavaScript in one file. No framework, build step, account, or server.
- Store all data in localStorage under one key, with a version number.
- Store every amount as integer cents, never as floating-point dollars.

Data model:
- Group: id, name, currency, members [{id, name}], archived.
- Expense: id, groupId, description, amountCents, paidBy [{memberId, cents}], splitType (equal | exact | percent), shares [{memberId, cents}], date, category, note.
- Payment: id, groupId, fromMemberId, toMemberId, amountCents, date.

Features:
- Create groups and add members by name.
- Add expenses with one or more payers.
- Group balance screen: who is owed and who owes.
- "Settle up" screen that lists the fewest transfers needed to clear all balances.
- Record a repayment.
- "Copy summary" button that produces plain text: each person's balance and the list of transfers.

Edge cases the build must handle correctly:
1. Equal splits that don't divide evenly assign the leftover cents one at a time, starting with the first member, so shares sum exactly to the total.
2. Percent splits must total 100%. Exact splits must total the expense amount. Block saving otherwise and show the difference.
3. Multiple payers: the paid amounts must sum to the expense total.
4. Every group's balances always sum to exactly 0 cents. Show a warning if they don't.
5. A member with a non-zero balance can't be removed until settled.
6. Editing or deleting an expense recalculates all balances.
7. Each group has one currency. No mixing currencies inside a group.
8. Negative amounts are rejected. Refunds are entered as a separate expense with a note.

Backup and import:
- One-click "Export JSON" and "Export CSV" per group. "Import JSON" restores a backup after confirmation.
- "Import Splitwise CSV": the file has Date, Description, Category, Cost, Currency, then one column per person holding (paid minus owed). Skip the final total row. Create members from the column headers.

Layout:
- Mobile-first, readable at 375px wide. Large tap targets for adding an expense at a restaurant table.

Exclusions:
- No publishing, App Store submission, accounts, login, payments, payment-app links, notifications, or social features.

Write a short README explaining where the data is stored, how to export and restore it, and how to share summaries.

Keep it under 900 lines. This is a personal tool, not a product.

What is the short answer?

Yes, with one catch. AI can build a bill-splitting app that replaces Splitwise Pro, about $4.99 per month in the US, for your trips and shared household costs. The catch is that you become the group’s bookkeeper.

Splitwise works because everyone logs in and adds their own expenses. A personal web app keeps data in your browser only, so you enter every expense and share a text summary of who owes whom. For a couple, a flat of 3, or a single trip, that works well. For a group of 8 who all pay for things, it gets tedious.

Option	Best for	Upfront cost	Ongoing cost	Main limitation
Build a rough prototype	People splitting one trip’s costs who want to test the idea	$0 and 3–4 hours	$0	Rounding and multi-payer expenses untested
Build a daily-use personal version	Couples or small flats where one person already tracks shared costs	$0–$20 and 8–10 hours	$0	Only you enter expenses; others see a pasted summary
Keep paying for Splitwise Pro	Groups where everyone adds their own expenses and wants receipt scanning	$0	$4.99/mo or about $29.99–$39.99/yr	Monthly cost for features the free tier partly covers

The ready-to-paste starter prompt at the end of this page builds the daily-use version.

What should the first version include?

The first version needs correct balances and a settle-up list. If those two are wrong, nothing else matters.

• Groups with named members
• Expenses split equally, by exact amounts, or by percentage
• Balances per group and a settle-up screen with the fewest transfers
• A way to record a repayment
• A copy-to-clipboard summary for the group chat
• Deliberate exclusion: no accounts, no shared login, no receipt scanning. Those need a server or paid APIs.

What can AI build reliably?

ChatGPT and Claude build the core of a splitter reliably: groups, members, expense forms, and a balance table. The debt simplification step, turning a web of debts into the fewest payments, is a well-known algorithm that AI writes correctly on the first try.

Concrete things AI gets right:

• Equal, exact, and percentage split forms with live validation
• Balance cards that show “Anna owes you $23.40”
• A settle-up list with 2 or 3 transfers instead of 6
• A text summary formatted for WhatsApp or iMessage

Where will AI need human help?

Rounding is the edge case that breaks splitters. Split $10 three ways and you get $3.33, $3.33, $3.33, with 1 cent missing. If the app stores decimal dollars, balances drift by a few cents over a month and never settle to zero. The starter prompt forces integer cents and assigns leftover cents explicitly. Test it anyway.

Other checks to run:

• An expense paid by 2 people
• Editing an old expense and confirming every balance updates
• Removing a member who still owes money

The decision only you can make is how your group shares the ledger. Agree up front that you’re the keeper, and that everyone sends you their receipts. That social decision matters more than any feature.

Privacy is simple. Balances stay on your device. The summaries you paste into chats are the only data leaving it.

Can I get my data out of Splitwise?

Yes. The free version exports each group to CSV, and Splitwise Pro adds full JSON backups downloadable from the Splitwise website.

The trap is the CSV format. As a Spliit GitHub issue documents, each person’s column holds what they paid minus what they owed for that expense, not their share. The last row is a total per person. An importer that reads those columns as shares produces wrong balances.

The starter prompt’s importer handles this format. After import, compare each person’s final balance with what Splitwise showed before you cancel. If they match to the cent, you’re done.

Export every group, including settled ones, if you want the history. Comments, receipt photos, and edit history don’t come through in the CSV [VERIFY].

How long will it take and what will it cost?

Rough prototype: 3–4 hours and $0 on a free AI plan. Groups, equal splits, and balances.

Dependable daily-use version: 8–10 hours across a week, including unequal splits, multiple payers, the Splitwise importer, and rounding tests. $0–$20 for one optional AI month.

Splitwise Pro costs $4.99 per month in the US. A $20 AI month pays for itself in about 4 months. On an annual plan around $30–$40, it takes 6–8 months.

These are estimates from reviewing the feature set, not measurements from a completed build.

Which AI tool or approach should I use?

Claude is a strong default because you can preview the app in the chat and test a 3-way split immediately.

ChatGPT works equally well and helps you write test cases, like “split $100 between 3 people where one pays $60”.

Cursor helps once the file is large and you want to fix the settle-up logic without touching the forms.

Lovable makes it easy to add shared logins so friends add their own expenses. That’s appealing, but it turns a personal tool into a hosted app with accounts to maintain.

What will I need to maintain?

No server, API, or bill. Your recurring work:

• Export a JSON backup after every trip or at the end of each month
• Paste the summary into the group chat when it’s time to settle
• Archive finished trips so the home screen stays clean

If you want the app on your phone at dinner, host the file on a free static host and add it to your home screen. Your data still lives only in that phone’s browser.

Should I build it, buy it, or reduce scope?

Build it if you already act as the money person in a couple, flat, or trip, and one ledger that you control suits the group.

Buy it if everyone in the group wants to add their own expenses from their own phone, or you rely on receipt scanning and currency conversion. For shared trips across currencies with 6 or more people, Splitwise earns its price.

Reduce scope if you split costs with one partner. Skip groups entirely and build a 2-person ledger with one running balance. That’s under an hour of work.

Recommendation: build it for couples and small flats where one person handles the money. Keep Splitwise, even on the free tier, for big friend groups where shared entry is the whole point. If shared costs are part of a household budget, track them inside the YNAB-style budget build instead.

What if I wanted to ship this to other people?

Shipping this means accounts, a shared real-time database, invitations, and conflict handling when 2 people edit the same expense. That’s a different, much bigger project than a personal ledger.

Starter build specification

• User: You, for your own use, as the bookkeeper for your group.
• Problem: Paying $4.99 per month for Splitwise Pro to track who owes whom.
• Core workflow: Add an expense after dinner, see updated balances, copy the settle-up summary into the group chat.
• Required features: Groups, 3 split types, multiple payers, balances, simplified settle-up, repayments, text summary.
• Deliberate exclusions: No accounts, no multi-user editing, no receipt scanning, no store submission.
• Data ownership: Stored in localStorage as integer cents. Exported as JSON and CSV. Imported from Splitwise group CSVs.
• Definition of done: You split $10 three ways and balances sum to exactly $0.00. An imported Splitwise group shows the same final balances Splitwise did.

If you split project costs with clients, that’s an invoicing job; see the Invoice Ninja build. For scoping help, read can AI build an app for me.

Sources and verification

• ShareCost vs Splitwise (BudgetBakers): Splitwise Pro at $4.99 per month, roughly $29.99–$39.99 per year by region, as of July 2026 [VERIFY on splitwise.com/pro].
• Splitwise on the App Store: Pro features (receipt OCR, JSON backups, currency conversion) and CSV export in the free app.
• Spliit GitHub issue on Splitwise CSV import: Splitwise CSV column format.

This app was not built and tested for this page. The verdict is based on Splitwise’s store listing, third-party documentation, and a review of the feature set. Splitwise’s price was checked on September 24, 2026.

Reviewed on September 24, 2026.