SumBudget
A household budget built around one question: what can we spend before the next payday? It reads the bank every day, sorts spending, finds the bills and paydays, and can add item-level detail from grocery and Amazon receipts.
Working and tested on one computer. Not online yet.How close is it to release?
- FoundationsDone
- Server and budget logicDone
- WebsiteBuilt · polishing
- Online for a few testersNext
- Phone appsLater
- Public releaseLater
The website and its server work end to end on the owner's computer: create a household, connect a bank, answer three setup questions, and use every screen. What's missing is mostly outside the code: choosing where it's hosted, a few privacy steps that have to happen before anyone else's bank data goes in, and the owner reviewing the latest changes. There's no date yet; the owner sets it once hosting is chosen.
What stands between now and a first outside tester
In rough order. Most of these are decisions or sign-offs only the owner can make.
- Review and approve the latest workPrivacy features, two-step sign-in and two rounds of design fixes are finished and tested, waiting for the owner's hands-on check.Owner
- Web address bought: sumbudget.appRegistered October 1, 2026. Until the website moves in, it shows this status page, updated with every change.
- Choose hostingWhere the website and its database run. Everything else online depends on this.Decision
- Prove household walls on the real databaseTests already show one household can't see another's data; this repeats the check on the production database itself.To do
- Pick an email serviceNeeded for password-reset and lost-phone links. Until then the online version sends no email.Decision
- Privacy policy sign-offA draft exists and shows on the site. It still needs a dedicated contact inbox and a lawyer's review.Owner
- Choose the bank connection for testersSimpleFIN works today and suits a small beta; Plaid is the likely choice for a public release.Decision
- Set it up online and invite testersThere is no public sign-up. The owner creates each household's login, and testers set up from there.Owner
What works today
Setup in four steps
Sign in, connect a bank (or start by hand), say who's in the household, confirm the pay schedule it finds in your paychecks.
A Home you arrange
Left until payday, safe to spend today, what's coming up, what to review. Each person picks, orders and sizes their own cards.
Transactions
Search, filter by month, group, category or account, sort, and download the list. Fix a category once and it can remember the store.
Bills and paydays
The next 30 days of expected bills and paydays, plus everything that repeats, with its cost per month and per year.
Budget and spending review
A proposed budget from your own history, progress by category, and a plain-language review of where each 100 dollars goes.
Receipts
Item-level detail from grocery receipts and Amazon orders, matched to the bank charge, so "groceries" can be broken down.
Private mode
Hides every amount on every screen, charts included, so the app can be open in front of other people.
Look around first
A made-up sample household to click through without signing up or connecting anything.
Light and dark, your colors
Theme, accent color, density, whole dollars or cents, and which tabs show on a phone.
Privacy and security
This handles bank data, so privacy came before features at each step. Built and tested so far:
- Each household's data is walled off from every other, checked on every route
- Two-step sign-in for everyone, with recovery codes, and a check against known-leaked passwords
- Bank access is stored encrypted, and its key can be rotated
- Collects less: names, loyalty numbers, addresses and full card numbers are dropped from receipts before saving
- Download all your data, or delete the whole household, from the app
- The app never sees your bank password; the bank connection service handles it
- Same wall check on the real hosted database, policy review, privacy inboxThe steps listed above, before anyone else joins.Before testers
Current polish work
An independent design review of every screen, at phone and desktop size in light and dark, found 37 issues. Two rounds of fixes are done; the rest are small.
Now: making it feel premium
The website works but looks plain, so the current stretch is about craft. A research pass looked at the best-regarded budgeting apps and at products known for their design, and set the direction:
- Research and design directionA display typeface for the big numbers, depth from soft layers instead of outlines, calmer colors, short motion that respects "reduce motion", and two new ideas no other budgeting app has: a month calendar with a "pace" line, and a Private mode that still looks finished.
- Type, surfaces and a wordmarkNext
- A real month chart, and designed loading, empty and error screensTo do
- A fresh, independent design critique of the resultTo do
Every change runs through automated checks: the budget math against the original spreadsheet, the server, the screens, real-browser runs at phone and desktop width, accessibility, and a scan that blocks any key, password or personal detail from getting into the code.
After the first testers
- An optional AI helperAsk a question about your own budget in plain English and get an answer worked out from your real numbers, plus a short weekly check-in. Off unless a household turns it on, sends only what the answer needs, never anything about another household, and has daily and monthly limits. The plan is being written first; it will be built and tested with a stand-in before a real AI is connected.Planning
- Phone apps for iPhone and AndroidThe website is already built with the same toolkit, so most of it carries over.Later
- Receipts for everyoneToday the receipt capture works only for the owner's household; testers would get budgets without item detail at first.Later
- Pricing, a company, and the app storesNot decided yet.Decision