How to track expenses without linking your bank
You can keep an accurate month without giving any app your net banking login. The method has three parts — catch what is already being sent to you, import what is not, and spend a few minutes a week closing the gap.
The idea: you are already being sent the data
Every card swipe, UPI payment and account debit generates a message to you within seconds. Your bank sends it whether you want it or not. That stream is a nearly complete record of your spending, and it is already in your possession — it just arrives as prose instead of rows.
Tracking without bank access means turning that stream into a ledger, rather than asking a third party to fetch the same information from the source. You end up in about the same place, without anyone holding credentials to an account that can move money.
Step one: capture what arrives on its own
There are three ways in, and they stack.
SMS, on Android. The most complete, because it catches nearly everything. You opt in, then link the sender IDs your banks and payment apps use — the app can scan your inbox on the phone to suggest which ones, without uploading any message bodies. From then on, alerts that look like bank or payment messages become drafts.
Email forwarding. For anything that arrives as a receipt rather than an SMS — card statements, order confirmations, subscription renewals. Forward it and it becomes a draft.
Telegram. Useful for the thing you want to record in the moment: send a short message when you pay cash, and deal with it later.
Everything from all three arrives as a draft, not an entry. It waits in an inbox until you approve it, and dismissing one leaves your books untouched.
Step two: import what capture cannot see
Capture starts from the day you switch it on, so the months before it are missing, and any account whose alerts you do not receive stays invisible. Both are fixed the same way: download a statement and import it.
CSV, TSV, text and Excel files all work, and the columns are detected for you rather than needing a specific layout. It is worth doing once at the start for the previous two or three months — a record that begins with history is far more useful in week one than one that starts empty, because you immediately have something to compare against.
Step three: the weekly review
This is the part that makes the method work, and it takes about five minutes.
- Clear the drafts. Approve what is right, fix what is mis-categorised, dismiss the duplicates. Doing this weekly rather than monthly is the difference between recognising a merchant name and guessing at it.
- Add the cash. Nobody sends an alert for cash. Reconstructing one week is realistic; reconstructing a month is not, which is the real argument for the weekly rhythm.
- Check the categories that moved. Not to feel bad about them — to notice which ones you did not expect. That is the entire return on the exercise.
How accurate this actually gets
Being straight about the gaps, because they are predictable:
- Cash is on you. Always, with any method. If most of your spending is cash, the weekly review is the method, and capture is a small helper.
- A bank you have not linked is silent. Until you add its sender IDs, its alerts are not recognised.
- Some messages are not parseable. Unusual formats come through as drafts you complete by hand, which is still faster than typing from scratch.
In exchange you get something a linked account does not give you: every entry has passed under your eyes once. People who track this way tend to know what their money did, rather than merely having a record of it. If you are weighing the two approaches, the full comparison is here.