Bank Statement to Tally
Built for CA firms and bookkeepers
Bank statement PDFs turned into verified, Tally-ready books in minutes.
| Date | Narration | Balance | |
|---|---|---|---|
| 02 Sep | UPI/SWIGGY/4218… | 4,82,116.40 | ✓ |
| 02 Sep | NEFT/MEHTA TRADERS | 5,68,616.40 | ✓ |
| 03 Sep | IMPS/OWN A/C ICICI | 4,18,616.40 | ✓ |
| 04 Sep | ACH/AWS INDIA | 4,00,214.29 | ✓ |
| 05 Sep | UPI/RAKESH KUMAR | 3,98,214.29 | ✓ |
| 05 Sep | RTGS/SUNRISE DIST. | 6,38,214.29 | ✓ |
Overview
Every bank lays out its statement differently, and retyping them into Tally is slow and error-prone. This tool reads the PDF, works out the layout itself and refuses to produce output unless every running balance and total matches the bank. Then it groups transactions by party so the accountant codes a whole party in one click, and remembers the choice next month.
Before and after
- Statements arrive as PDFs, often locked, in a different layout per bank.
- Rows were copied into Excel and then Tally, with digits transposed or rows missed.
- Each of thousands of rows needs the right ledger and voucher type.
- Most converters guess on messy PDFs and quietly produce wrong numbers.
- A statement engine that detects columns, date and amount formats from the file itself.
- A balance-chain check on every row, and a total check against the bank’s printed figures.
- A coding screen that groups transactions by party and category for one-click ledger assignment.
- A memory of ledger choices, so next month’s statement arrives pre-coded.
How it works
- 1
Read
The PDF is read and the bank’s layout detected automatically.
- 2
Verify
Previous balance + amount = new balance, on every single row.
- 3
Group
UPI, NEFT and IMPS narrations are parsed into parties.
- 4
Code
One ledger choice codes all of a party’s rows.
- 5
Export
A Tally-ready workbook, with doubtful rows held for review.
What it does
No retyping errors
If any row’s balance breaks, conversion stops with a clear reason.
Works with most banks
Profiles for HDFC, ICICI, SBI, Axis, Kotak and Union Bank, plus a general reader.
Gets faster every month
Party-to-ledger choices are remembered across statements.
Fewer wrongly filed entries
Own-account transfers recognised by exact name match, never guessed.
Every decision explained
Each row carries the rule, reason and confidence behind it.
Locked PDFs handled
Password-protected statements open; scanned images are refused rather than misread.
Under the hood
- Three table-extraction strategies, including a fix for statements whose column lines don’t meet.
- Date ambiguity (5 Aug or 8 May?) resolved by testing formats across the whole file.
- An early rule mis-filed 271 transactions as transfers; replaced with exact name matching.
- 220 automated tests, including real statements from several banks.
Outcome
Related work
Tally API Gateway
A modern front door for TallyPrime, so stores, apps, spreadsheets and AI assistants can post entries and pull reports.
Invoice AI for Tally
Upload a purchase invoice, review what the AI read, and post it into Tally in one click.
Want this for your business?
Tell me how your team works today. I’ll tell you how much of it a system like Bank Statement to Tally could take over.