How It's Going

The goal from the start was to add AI without making it the point. WellSpnt is a manual tracking app, and that isn't changing. But there are moments where a little context or a quick answer makes the experience meaningfully better, and that's where these features live.

The Bucket suggestion was the first one built, and it set the tone for the rest. The logic for categorizing a budget item lives in the app's own code as structured rules, not inside a prompt. The AI handles the explanation. That separation keeps the suggestions consistent and makes genuinely ambiguous cases honest about the ambiguity, rather than just picking a side.

Something else got removed this update too. The Close Month feature had been in the app since early on, but a closer look at what it actually did revealed a problem: it performed an irreversible balance sweep with no real safeguards. More importantly, that kind of automatic sweeping doesn't fit how WellSpnt thinks about money. Removing it was cleaner than trying to fix it.

A lot of small decisions came together this stretch around the idea of context. The page taglines (Budget/"The Plan", Bank Accounts/"The Reality", and so on) weren't about being clever. They were about giving each page a reason to exist before you've used it. The Getting Started checklist followed the same logic: not a tutorial, just a quiet nudge toward the right first steps.

The shared account picker was the most satisfying fix. Every place in the app where you choose an account, adding income, logging a spend, editing a transaction, was slightly different. Unifying them onto one component with consistent ordering and colored type pills was the kind of cleanup that doesn't feel like much until it's done, and then it's hard to imagine it the old way.

There were also several real bugs fixed this stretch that had been hiding behind surface-level symptoms. Popovers clipping, closing, or stretching unexpectedly. Decimal columns falling out of alignment. The fixes weren't glamorous but they made the app feel more solid.

Bank Accounts came together in one session, which surprised me. The architecture question that took the longest to resolve was actually a small one: what does a credit card account mean in an app that already has a Debt page? The answer was that it shouldn't hold its own balance. It should just link back to the debt record that already exists. Once that clicked, everything else followed.

A lot of this session was also about replacing modals with inline interactions. Hiding a budget item, editing a future account, reordering the budget list. All of these used to open a separate modal. Now they happen directly on the page. It's a subtle shift but it makes the app feel faster and more direct to use.

Calling this v1.0 felt right not because the app is finished, but because it finally feels like something you could hand to another person and have them understand it.

The app had outgrown a single file. At around 2,400 lines, app.py had become hard to navigate and harder to hand off cleanly. Breaking it into smaller focused modules was overdue, and doing it carefully meant catching three real bugs that had been quietly sitting in the code the whole time.

Getting on GitHub the same week felt like the right moment. The app finally had a structure worth preserving, and a private repo means there's a real record of how it's been built.

The visual side got the same treatment. Starting with the Home page and side nav, then working through every remaining page, the whole app now shares one consistent visual language. It's the first time it's felt that way from top to bottom.

A recurring pattern this stretch was finding things that didn't quite match and bringing them in line. Forms that looked different from each other. Dropdowns rendering with default browser styling. Modals holding onto stale data. None of it is dramatic, but fixing it made the app feel more like one thing.

Money Owed to Me also found its right home. It had been sitting in Future Accounts, which never made sense. It's not a savings goal. It's more like the mirror image of debt. Moving it to the Debt page cleaned up the architecture without needing to build anything new.

The Settings page is just getting started. Themes, a scroll preference, CSV export. There's more to add over time, but having the foundation in place means future options have somewhere to live.

Debt had been easy to put off because it required real decisions. Once I committed to treating it as its own group with its own section on the budget page, the rest followed naturally.

The hardest part was getting the Net calculation right. Contributions to savings reduce your net. Withdrawals show up as income. That logic sounds simple until you walk through a real scenario, and walking through those edge cases before they became bugs was worth the time.

Future Accounts also got a full rebuild. Getting it right meant making sure Debt, Trips, and Future Accounts all follow the same pattern: create once on their own page, reflect automatically on the budget page. That consistency is what made the app start to feel like a system.

Real life doesn't follow a clean monthly budget. Bills get skipped. Travel spending doesn't fit neatly into a category. Income shows up on a different day than expected. This update was about making the app handle all of that honestly.

Skip ended up being the trickiest piece to get right. The first version would have skipped every occurrence of a weekly bill for the whole month, which isn't how anyone actually lives. Keeping it on the calendar and scoped to one date at a time made it feel correct.

Trips went from a vague idea to a full feature with its own page, transaction linking, and a clean list and detail view. It's the first time WellSpnt has had something that lives completely outside the monthly budget.