@can.chill.boh: Description: Stop manually typing out your bank transactions I built a system using Claude AI that automatically converts my Maybank & UOB PDF statements into a clean, categorised Excel file every month — with duplicate detection, auto-categories, and zero manual work. Full setup prompt in the comments #ClaudeAI #AI #PersonalFinance #FinanceTips #MalaysiaFinance

Can Chill Boh :

The prompt : I want to build a system that parses bank statement PDFs and merges them into a single combined Excel and CSV file using Python and pdfplumber. Setup: I will tell you which banks I have and upload one sample PDF per bank to start You will inspect each PDF, understand its layout, and build a dedicated parser for it Once all parsers are built, you will build a merge script that combines everything Scripts to build (one at a time as we go): One parser per bank — extracts transactions from PDF into structured data parse_utils.py — shared utility to clean transaction descriptions and extract any metadata fields that appear consistently (e.g. location, payment method, instalment progress) — based on what you find in my PDFs categories.py — keyword-based category and sub-category mapping, built based on the actual merchants found in my statements. I will review and correct the suggested categories before finalising. merge_all.py — merges all banks into one combined sheet with CSV persistence for appending across sessions Formatting rules: No title row, no totals row — just headers and data Auto-filter and freeze header row Deduplication: Statement-level: Bank + Card/Account + Statement Date If a duplicate statement is detected, skip all its rows and flag it in chat Legitimate repeated transactions within the same statement must be preserved CSV persistence: merge_all.py loads an existing CSV as the base if one is provided New PDFs are parsed and appended on top Deduplication runs across existing and new data combined Both CSV and Excel are saved at the end How it works going forward: Each session I upload my new PDFs plus my existing CSV The script appends new statements and skips any already in the CSV I can upload all banks at once or one at a time — both work To start: I will upload one sample PDF from each of my banks and tell you the bank name. You inspect each one, confirm the format, build the parser, and we test it before moving on to the next bank.