Upload your statement PDF. See where your money goes β spending, categories, unusual charges. Private by design: we analyse and forget.
Tap to choose your statement
or drag & drop the PDF here
π Locked PDF? Unlock it on the next step.π‘ Drop 2+ statements to merge them.
What you'll get
π
PrivateAnalysed in memory, deleted instantly
β‘
FastResults in a few seconds
π¦
Multi-bank mergingAuto-detects and combines HDFC, SBI, Axis & more
π
Reading your statementβ¦
π This is sample data β upload your own statement to see your real numbers.
π§ͺ This bank isn't officially supported yet. Transactions below were extracted automatically and the running balance reconciled correctly, but this format hasn't been human-verified the way our supported banks have β please double-check totals before relying on this.
β οΈ We couldn't fully verify this extraction. The running balance in your statement doesn't add up against what we parsed (off by β first shows up around ) β a page or transaction may have been missed. Numbers below could be incomplete.
Net total
βΉ0
+βΉ0Β·ββΉ0
Excel Β· Tally XML Β· QuickBooks/Zoho CSV β one purchase, every format
NetβΉ0
Transactions0
Where it went
βΉ0
Total spend
Tap a category to show or hide it
Top places you paid
π This month vs last
Only one month in this statement, so there's nothing to compare it against yet. Drop in a second month's PDF β or both at once β to see what changed.
Money out
βΉ0
Money in
βΉ0
Net
βΉ0
What moved the most
Category spending was flat between these two months.
Month by month
β οΈ Unusually large payments
These are much bigger than your typical spend β worth a quick look.
β Nothing stood out. Every payment here is in line with how you normally spend.
π Recurring subscriptions
Same amount, roughly every month β the auto-renewing charges easiest to forget about.
β No auto-renewing charges found. Nothing here is quietly billing you every month.
If you cancelled every one of these, you'd keepβΉ0a year
βΉ0 a month across 0
All transactions
Wrong category? Tap it to fix β the change sticks for that merchant, on this device.
Date
Merchant
Category
Amount
Choose what to download
One purchase unlocks every format. Pick what you need now β you can reopen this and grab the rest without paying again.
Paid once per statement Β· secure Razorpay checkout Β· we never see your card details
What is SpendPattern?
In plain English
A bank statement is a wall of cryptic text β UPI-SWIGGY-swiggy@ybl tells you almost nothing at a glance. SpendPattern reads that wall for you and turns it into a picture: how much went out, how much came in, what you spent it on, and which charges look unusual.
You upload your statement PDF, wait a few seconds, and get a dashboard. No account to create, no forms to fill in, nothing to install.
Step by step
Upload. Drop in your statement PDF. Locked with a password? You'll be asked for it β it's used to open the file and then discarded.
We read it. Your bank is auto-detected, and every transaction is pulled out with its date, description and amount.
We sort it. Each transaction is matched to a category β food, transport, bills, shopping and so on β from the merchant name in the description.
You explore. Charts, top merchants, recurring subscriptions, and unusually large payments. Tap any category to filter, or search for a specific merchant.
Optionally, you export. A paid one-off download gives you the same data as an Excel report, a Tally XML import, or an accounting CSV.
What happens to your data
Your PDF is processed in memory and never written to disk on our side.
Nothing about your statement is stored on our servers after the response is sent β there is no database of your transactions.
"Past analyses" lives in your browser's own storage, on this device only. It keeps a small summary, never the original PDF. Clear it any time from that panel β worth doing on a shared computer.
Payments go through Razorpay's own hosted checkout. Your card or UPI details are entered on their page, not ours, and we never receive them.
For the technically minded
Text is extracted from the PDF and parsed with per-bank layout rules β currently HDFC, SBI, Axis, PNB, Indian Overseas, City Union and Bank of Baroda, auto-detected from the document itself.
Categorisation is a deterministic keyword match against a merchant dictionary, not a language model. The same statement always produces the same categories, and nothing is sent to any AI provider.
"Unusually large" is a statistical outlier check against your own spending distribution, not a fixed rupee threshold β so it adapts to how much you normally spend.
Recurring subscriptions are detected by finding near-identical amounts to the same merchant at roughly monthly intervals.
Drop in several statements at once and they're merged into a single timeline, deduplicated where the date ranges overlap.
Where it can be wrong
A statement from an unsupported bank, or an unusual layout from a supported one, may extract partially or not at all. Nothing is silently guessed β you'll see it in the numbers.
A merchant that isn't in the keyword dictionary lands in a general category rather than being invented into a specific one.
Scanned or image-only PDFs have no extractable text and won't work; you need the digital statement your bank issues.
π¬ Feedback
Found a bug, or have an idea to make this better? Tell me directly β I read every one of these.
π³ Picked up from a shake β device/browser details are included below, feel free to describe what went wrong.
π A masked sample of your statement's text (account numbers and UPI handles starred out) will be included below, to help add support for your bank.
π³ Shake to report a bug
Optional. Lets you shake your phone to open this form from anywhere in the app. Your phone will ask permission for βMotion & Orientationβ β thatβs the iOS name for the shake sensor. Nothing is recorded or sent.