DEV Community: daniel zheng

Bill import without a bank login: parsing statements you already own

daniel zheng — Mon, 02 Nov 2026 01:51:57 +0000

Disclosure: Written by the developer of Money Tracker (Android: Money Tracker-Expense & Budget; iOS: Money Tracker:Expense & Budget). This is a technical/product note on bill and statement import with confirm-before-save — not a bank-aggregator review, not open-banking marketing, and not a claim that we log into your bank.

Expense apps love a clean demo: connect a bank, wait a few seconds, watch the month fill itself. The engineering story underneath is usually credentials, tokens, and a sync job you do not see.

I build Money Tracker. We still refuse that story as the default. Capture is typing, AI voice, receipt photos, and — on Premium — import from bills or statements you already have, turned into drafts you must confirm before they become history.

This piece is about that import path: what it is for, what it is not, and why “parse a file I exported” is a different trust model from “here is my bank password so you can pull forever.”

Import is not sync

Bank sync means the app keeps a standing relationship with your accounts: credentials or OAuth, recurring pulls, silent inserts.

Bill import (as we mean it) means: you already obtained a PDF, CSV, or similar artifact from your bank, card issuer, or biller — outside our app — and you hand that file to Money Tracker for a one-shot parse into candidate lines.

Same screen language as voice and OCR:

The model or parser may draft. A human edits before save.

Wrong merchant name on a statement line? Fix it. Fee vs purchase split oddly? Fix it. Then confirm. Nothing lands in the permanent book because a parser felt confident.

We do not ask for bank logins to make import work. If a workflow needs you to paste banking credentials into Money Tracker, that workflow does not exist in this product.

Why bother importing at all

Manual entry is honest and slow. Voice and receipt OCR shrink friction for today’s coffee and today’s paper slip. They do not help much when you already have a monthly card PDF and a weekend of catch-up.

Import exists for the catch-up case:

You exported a statement because the bank already authenticated you in their app or site.
You want those lines in your book without retyping twenty merchants.
You still want a checkpoint so a parser glitch does not invent history.

That is a narrower promise than “set and forget open banking.” It is also a promise we can keep without expanding credential surface into a budgeting tool.

What we will and will not claim

Will claim

Import turns supported bill/statement inputs into editable drafts.
Save is confirm-before-save, same family of rules as AI voice and receipt OCR.
Free tier still tracks expenses without bank linking; import extras follow in-app freemium (Premium), not a blog SLA.
Exact formats and quotas live in the product UI — posts go stale; the app does not.

Will not claim

That Money Tracker logs into your bank, scrapes sessions, or maintains open-banking connections.
That every bank’s PDF/CSV on earth parses perfectly.
That import replaces judgment about what “counts” as a personal expense (transfers, payments to yourself, refunds — you still decide).

If verified multi-bank aggregation is the product you want, use a tool built for that. Import here is intentionally boring: your file → our draft → your confirm.

A practical catch-up loop

In the bank or issuer app, export the period you care about (their auth, their export).
In Money Tracker, start import and pick the file.
Review draft lines: amount, date, payee-ish text, category guess.
Edit the weird ones. Drop the ones that are not expenses you want in this book.
Confirm. Only then do lines become history.

Treat the first import as calibration, not as a one-click autobiography. Parsers stumble on multi-line descriptions and local fee labels. The checkpoint is the feature.

How this fits freemium without a credential hook

Free Money Tracker runs with ads and supports everyday capture. Premium removes ads and unlocks deeper helpers — including bill import among other extras. The free path is not a trap that only becomes usable after you link a bank.

That matters for trust writing. “Free” that is actually “free until you paste banking identity” is a different product than freemium ads you can see.

Closing

Bill import without a bank login is not magic. It is a deliberate smaller API: parse what you already own, draft loudly, save only on confirm.

If you want silent sync, we are the wrong app. If you want a book that stays on your side of the credential fence — with voice, OCR, and import as drafts — that is the line we are willing to defend in public.

Product site (EN root):

https://money.quhou123.com/?utm_source=devto&utm_medium=article&utm_campaign=offsite_w1

Play (optional):

https://play.google.com/store/apps/details?id=com.freeman.moneymanager

I help build Money Tracker. This draft is about bill/statement import with confirm-before-save — not bank login, not open-banking aggregation.

Why freemium ads and confirm-before-save beat silent bank sync for trust

daniel zheng — Thu, 29 Oct 2026 01:40:36 +0000

Disclosure: Written by the developer of Money Tracker (Android: Money Tracker-Expense & Budget; iOS: Money Tracker:Expense & Budget). This is a product-trust note — freemium ads, confirm-before-save, and why we refuse silent bank sync — not a bank-aggregator review and not a launch hype post.

Expense apps often sell trust as a shortcut: connect your bank once, and the ledger fills itself. The pitch sounds caring. The trade is not small. You expand the credential surface so the app can claim “zero effort.”

I build Money Tracker. We took the opposite bet on purpose: keep ads visible on the free path, keep every AI draft editable before it becomes history, and never ask for a bank login to make tracking feel automatic.

That is not nostalgia for typing. It is a trust model. Ads you can see — and Premium you can buy to remove them — are an honest deal. A silent bank sync that “just works” is a different deal: convenience paid for with identity.

This piece is about that choice. Not trip AA books, not category budget alerts, not household share invites. Those stories exist elsewhere. Here the subject is why freemium + confirm-before-save is the product line we are willing to defend in public.

Trust is a surface-area problem

When someone installs a money app, they are not only buying categories and charts. They are deciding how wide the blast radius gets if something goes wrong: a wrong category, a wrong amount, a wrong “who paid,” or worse — a credential reused beyond the bank’s own app.

Silent bank sync shrinks typing and expands delegation. The app becomes the authority over what happened. That can be fine for people who want open-banking style aggregation as the center of the product. It is the wrong default for anyone who will not paste bank passwords into a budgeting tool “so the bars update.”

Money Tracker does not link bank accounts and never asks for bank logins. Everyday tracking, categories, budgets, and reports work without that expansion. Capture stays on your side of the fence.

If verified multi-bank sync is your core workflow, pick a tool built for that. We would rather stay clear than blur the line in a trust essay.

What “confirm-before-save” actually means

AI in an expense app is useful when it drafts. It is dangerous when it commits.

In Money Tracker the rule is the same on every AI path:

The model may draft. A human edits before save.

That covers:

AI voice — “lunch 12,” “taxi 45” — draft, then confirm.
Receipt photo — OCR suggests a total and merchant-ish text — draft, then confirm.
Android Smart Notification Entry (yearly Premium) — payment-style notifications can become draft candidates — still save only after you confirm. It is not a silent bank API. iOS does not get that notification path; type / voice / receipt cover capture instead.

A wrong coffee in a personal book is annoying. A wrong line that landed because a demo skipped the checkpoint is how trust dies — with the product, and sometimes with the people sharing a book with you.

Confirm-before-save is slower than “auto-import everything.” It is also the reason we can talk about AI without pretending the model owns your history.

Why visible ads beat hidden credential trades

Freemium is easy to abuse in marketing. “Free forever” with a vague upsell. “Free” that is actually a lead-gen funnel into bank linking. We try to keep the deal boring and readable:

The app is free with ads.
Core tracking, categories, budgets, and reports work on the free path.
Yearly Premium removes ads and unlocks extras such as unlimited AI entry (after free trial credits), bill import, app lock, more books and accounts, team / share extras (initiating invites; accepting stays free), Android notification entry (confirm-to-save), and choosing your budget month start day.
Exact quotas follow in-app prompts, not blog posts.

Ads are not elegant. They are honest. You see what you are paying with (attention) unless you choose to pay with money. Silent bank sync as the free “hook” asks you to pay with identity and access before you have decided the product deserves it.

I would rather someone bounce after seeing an ad than feel tricked into expanding bank credentials to make a demo chart look full.

A practical week on this trust model

Here is the boring loop that matches the product shape — no bank password required:

Install, create a book, set one or two categories you actually care about. Do not invent twenty budgets on day one.
Log the same day. Type when you have ten quiet seconds. Voice on the walk. Receipt after a sit-down meal.
Edit before save whenever AI drafts. Wrong category → fix. Wrong amount → fix. Then save.
Optional (Android + Premium): turn on notification drafts if you want less typing after wallet / QR payments — still confirm; still not a bank login.
Review mid-week, not only on statement day. The ledger only teaches you if it has real lines.

What this is not: automatic fairness, open banking for every roommate, or a guarantee that OCR works in every noisy café. Social and capture habits still matter. The product job is to make the honest path lighter than a spreadsheet nobody updates — without selling trust as a login form.

Who this is for (and who should skip it)

Worth a look if you:

want an expense and budget app on Android or iPhone without pasting bank passwords,
prefer edit-before-save over silent auto-import,
are fine with free + ads, or optional yearly Premium when you want extras,
care more about a clear credential boundary than about “zero effort” marketing.

Skip if you need certified multi-bank sync as the center of the product, want zero confirmation steps, or only want a desktop spreadsheet after the month ends.

What I will not claim

I do not claim ads make anyone happy. I do not claim confirm-before-save is as fast as silent sync. I do not claim voice or OCR works in every language and lighting condition. Blog posts age; in-app prompts and the store listing do not owe us marketing fiction. No awards invented here.

What I will claim: freemium ads you can see, drafts you must confirm, and bank logins that stay in the bank app — a quieter trust model for people who refuse to expand the credential surface just to fill a ledger.

Product overview (same boundaries):

https://money.quhou123.com/en/?utm_source=devto&utm_medium=article&utm_campaign=offsite_w1

I am daniel zheng, developer of Money Tracker. This piece is about freemium honesty and confirm-before-save as a trust boundary — including Android notification drafts and freemium limits that follow in-app prompts.

Shared household expenses without merging anyone's bank login

daniel zheng — Mon, 26 Oct 2026 01:53:36 +0000

Roommates, couples, and small teams hit the same awkward wall: everyone needs a shared picture of “what we spent,” but nobody wants to hand a budgeting app their bank password — or merge three personal logins into one “household account.”

I build Money Tracker. We kept drawing that line on purpose: share a book, not a bank identity.

This is not another “connect your bank and magic happens” post. It is about a quieter product shape that works for household pots, AA dinners, and team books when the hard requirement is: capture stays on your side of the fence.

The household problem is social, not “more OCR”

Solo expense tracking fails when you skip entries. Shared tracking fails for a different reason: unclear ownership of the ledger.

Someone paid the grocery run. Someone else covered the internet bill. A third person settled last month’s dinner and never logged it. Spreadsheets die after week two. Group chats become archaeology. And the “easy” product answer — link every card so the app owns the truth — asks people to expand the credential surface just to stop arguing about who owes what.

Plenty of households want less friction. That does not mean they want one shared bank login.

So the job we optimized for looks like this:

Keep a shared book (or an AA / split book) that everyone who needs it can see.
Log lines with typing, AI voice, or a receipt photo — then edit before save.
Settle later with notes that still make sense to a human.
Never ask for a bank password to make the sharing feel “automatic.”

If verified multi-bank sync is your core workflow, pick a tool built for that. We would rather stay clear than blur the line in a household post.

Books, not one blurry “Family” pile

Money Tracker is a freemium expense and budget app for Android and iPhone (store names: Money Tracker-Expense & Budget / Money Tracker:Expense & Budget). The unit that matters for sharing is a book — a separate ledger you can keep for home, rent+utilities, a trip pot, a side project, or a roommate kitchen fund.

That sounds small until you have lived with one undifferentiated list called “Everything.” Rent, personal coffee, and a group dinner do not belong in the same mental bucket. Separate books keep the household pot readable without forcing everyone to dump their entire personal history into one shared feed.

For roommates and friends, AA / split books cover “we split this meal / this weekend.” For longer-running household or team pots, shared / team books are the shape you want when more than one person should keep logging into the same ledger.

Honesty on freemium (exact quotas follow in-app prompts, not blog posts):

The app is free with ads.
AA / split books can stay free to create in ordinary use.
Team books, initiating share invites, and smart settle are among the extras that need yearly Premium.
Premium also removes ads and unlocks things like unlimited AI entry (after free trial credits), bill import, app lock, more books/accounts, and — on Android only — Smart Notification Entry (draft from payment-style notifications; still saves only after you confirm).

Trust the in-app copy if a version changes a quota. Blog posts age; the store listing and the app do not owe us marketing fiction.

Capture that does not require merging logins

Shared books only work if people actually enter lines. Waiting until “someone reconciles the joint account” usually means nobody does.

In Money Tracker, capture paths stay local to the person holding the phone:

Type when you have ten quiet seconds.
AI voice when you are walking out of the store (“groceries 86, household”).
Receipt photo when the slip has the categories you care about.

On every AI path the rule is the same: the model may draft; a human edits before save. A wrong coffee is annoying. A wrong rent-adjacent transfer inside a shared book is how trust dies between roommates.

None of those paths require linking a bank account. Money Tracker does not link bank accounts and never asks for bank logins. On Android, the optional Premium notification path reads notification text from other apps and still waits for confirm — it is not a silent bank API, and iOS does not get that path at all. Household sharing does not depend on it.

A practical household setup that survives a Tuesday

Here is the boring setup I recommend when people ask “how do we start without connecting banks?”

Create a household book (or an AA book if it is mostly dinners and weekend trips). Keep personal day-to-day spend in a separate personal book if you want privacy from the group.
Agree on three categories that matter for the pot — e.g. groceries, utilities, shared transport. Do not invent twenty labels on day one.
Log the same day. Voice or receipt draft after the purchase; fix merchant/category; save. The person who paid logs it, with a short note if someone else will reimburse.
Review the book weekly, not in a fight. Shared ledgers fail when the first read is an accusation.
Settle with the app’s split / AA tools when you need them — and with clear notes when you do not. Exact settle UI labels change by version; the durable habit is “the book shows the story before anyone Venmos anger.”

What this is not: automatic fairness, open banking for every roommate, or a guarantee that your cousin will open the app. Social tools still need social habits. The product job is to make the honest path lighter than a spreadsheet nobody updates.

Why “no bank login” belongs in a sharing article

Privacy posts often sound abstract until you share money with other people.

Merging bank logins for a household app expands trust beyond the developer: you are also trusting whoever else gets invited into the same financial surface, plus whatever sync vendor sits in the middle. Some households accept that trade. Many do not — especially roommates, short-term flatshares, and small teams that only need a pot ledger for three months.

Our constraint is deliberate:

Share the book. Keep bank credentials in the bank app.

That is why Soft promotion for Money Tracker keeps repeating the same boundary across different articles: receipt capture, travel books, notification drafts, and now household sharing. The feature set changes with the story; the credential line does not.

Who this is for (and who should skip it)

Worth a look if you:

need a shared household or AA pot on phones you already use,
want voice / receipt drafts with edit-before-save,
refuse to paste bank passwords into a budgeting tool “for convenience,”
are fine with free + ads, or optional yearly Premium when you need team/share extras.

Skip if you need certified multi-bank sync as the center of the product, want zero confirmation steps, or only want a desktop spreadsheet after the month ends.

Try it

Official site (this article’s UTM):

https://money.quhou123.com/en/?utm_source=devto&utm_medium=article&utm_campaign=offsite_w1

I do not claim shared books replace couples therapy, or that every roommate will log on time. I do claim a clearer product shape: household expense sharing without merging anyone's bank login — books you can share, capture you can review, and bank passwords that stay where they belong.

I am daniel zheng, developer of Money Tracker. This piece is a practical note on shared / team books and AA splits without bank linking — freemium limits follow in-app prompts.

Why our expense app never asks for your bank password

daniel zheng — Fri, 23 Oct 2026 01:32:46 +0000

Open-banking fear is rational. Once an expense app asks for your bank username and password — or a deep “connect account” flow that feels the same — you are trusting that product with credentials you would not paste into a random form. Even when the pitch is convenience (“we sync everything”), the trade-off is clear: your ledger gains speed; your bank identity leaves the bank app.

Plenty of people still want less typing after every QR / UPI / Pix / wallet payment. Those two desires collide. So when we built Smart Notification Entry in Money Tracker, we started from a hard line:

The app does not connect to bank logins. It never asks for your bank password.

That is not a soft disclaimer buried in settings. It is the product shape.

What we do instead of bank login

Money Tracker is a bookkeeping and budget app (Android store name: Money Tracker-Expense & Budget; iOS: Money Tracker:Expense & Budget). Capture stays on your side of the fence:

You can type entries.
You can use AI voice or receipt photo entry (results are editable before save).
On Android, Premium users can optionally let the app watch transaction-style notifications from other apps, draft a line, and wait for you.

None of those paths require handing over bank credentials. If you need verified bank sync as the core workflow, this is not that product — and we would rather say that upfront than blur the line.

High-level: how Android notification bookkeeping works

Android exposes a system permission called notification access. With it, an app you choose can read the text of notifications other apps post — the same banners you already see when a payment succeeds.

Smart Notification Entry uses that surface, not a bank API and not a stored password:

You open Settings → Smart Notification Entry (Premium) and turn the feature on.
Android asks you to grant notification access (so Money Tracker can monitor other apps’ alerts) and the app’s own notification permission (so it can nudge you when something looks like a transaction).
When another app posts a transaction-like notification, Money Tracker can treat it as a draft candidate.
You review it — amount, category, and related fields — then confirm. Until you confirm, it is not meant to become permanent ledger history.

At a product level, the pipeline is: notification text → draft → human confirm → save. We are not documenting parser heuristics, vendor-specific reverse engineering, or a “supported bank list.” Payment apps and notification formats change; naming untested packages would be marketing fiction. The durable idea is simpler: reuse the alert you already got, then keep the human in the loop.

You can also put apps you do not want monitored on a blacklist. After that, those apps’ transaction notifications are ignored, and any pending drafts from them are cleared. Remove them from the blacklist later if you change your mind.

Confirm-to-save is the privacy feature

“Automatic expense tracking” demos love to skip the checkpoint. In a ledger, the checkpoint is the feature.

A wrong coffee is annoying. A wrong transfer amount — or a duplicate — is how trust dies. So the design default is:

Suspected transactions land in a pending / to-confirm flow (you can open from the nudge, or process items in a list).
You can confirm one by one or batch-confirm with “recognize and import.”
Suspected duplicates are meant to force a confirmation path rather than silently stacking copies.
Ignore is always available when the alert was not a spend you care about.

That is the same philosophy as AI voice and receipt entry elsewhere in the app: AI or automation may draft; you decide what becomes history. Speed without a silent write.

Important honesty for tech readers: this is not silent full-statement import, not open banking, and not “set it and forget every line forever.” It is a lower-friction capture path for people who already get payment alerts on Android and are willing to tap Confirm.

Limits (please read before you expect magic)

Boundary	What it means
Android only	iOS has no Smart Notification Entry. iOS users should use manual entry, AI voice, or receipt capture instead.
Premium	Notification bookkeeping is a membership feature. Free includes ads and core tracking; Premium removes ads and unlocks extras (including this on Android).
Confirm-oriented	Treat it as draft-then-confirm, not “the app owns your bank.”
Blacklist	You control which apps may be monitored.
No bank password	We do not ask for one. Do not enter bank credentials into any expense app that should not have them — including ours.

Other Premium examples that often get mixed into “privacy” questions: app lock (PIN / biometrics), more books and accounts, unlimited AI entry after free trial limits, bill import, team books / share invites, and smart settle on AA. AA / split books can stay free to create; team books, initiating share invites, and smart settle need Premium. Exact quotas can change by version — trust in-app copy over blog posts.

Short FAQ

Does Money Tracker log into my bank?

No. It does not use bank logins or bank passwords.

Is notification entry “fully automatic”?

No. On Android Premium it drafts from transaction-style notifications; you confirm before those drafts are treated as saved bookkeeping.

Can I stop it from reading a specific app?

Yes — blacklist that app.

Why not just connect open banking everywhere?

Because many users in our markets want faster capture without expanding the credential surface. Notification text + confirm is a deliberate smaller trust boundary.

Will every payment app’s alert parse perfectly?

No. Formats differ; noisy or non-transaction alerts exist; you can ignore or edit. We do not publish an unverified “works with every wallet” claim.

Who this is for

Worth a look if you pay from phone apps all day, already see payment notifications, refuse to type bank passwords into a budgeting tool, and will spend two seconds confirming a draft.

Skip the notification path if you are on iOS only, need certified bank-sync as the center of the product, or want zero confirmation steps. The rest of Money Tracker still works without that feature.

Product overview (same boundaries):

https://money.quhou123.com/en/?utm_source=devto&utm_medium=article&utm_campaign=offsite_w1

Stores: Google Play (Money Tracker-Expense & Budget) · App Store (Money Tracker:Expense & Budget).

I help build Money Tracker. This piece explains why we avoid bank-password flows and how Android notification drafts + confirm-to-save are meant to work — including the Premium / Android-only limits. Originally published on Hashnode; this is a cross-post with canonical URL set.