Analyse a bank statement with AI
Check monthly spending against the actual rows in a statement export, including recurring payments and possible duplicates.
Analyse a bank statement with AI when you want to understand a change in spending without typing transactions into a second spreadsheet. With txn-statement-analyzer, a local Python script calculates the figures and the agent explains the result. A possible duplicate or an unusually large charge comes with row references so you can return to the source.
Choose one account and currency per file. A report can show the whole statement or a particular date range. Recurring payments are assessed against the available history, so a wider export can make regular charges easier to recognise.
What it does
- Compare money in and money out by month.
- See which spending categories make up the largest share.
- Check same-day charges from the same merchant for a possible duplicate.
- Review subscriptions suggested by at least three roughly monthly payments.
- Find unusually high amounts within a category and compare them with receipts.
Install and use
- Use an authorised CSV export and an agent with local access and Python 3. A bank login is unnecessary.
- Identify the date, description and amount columns, or the separate debit and credit columns if your bank uses them.
- Ask, “Show my spending for September and include statement rows for every flagged charge.” Check the sign convention and any skipped rows.
FAQ
Source
Made by Dayesh Raval. Original skill: txn-statement-analyzer, published on GitHub under the MIT licence.
View on GitHub