Receipts to spreadsheet: check what the reader got, then download
This is a submission for the Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle What I Built A web page that turns receipt pictures into a spreadsheet you check before you use it. You drop in pictures, the page reads vendor, date, total and tax, it highlights anything it is unsure of, you fix the fields and tick "I checked this", then you download a CSV. No terminal, no account, no API…
This project is a web page that turns pictures of receipts into a spreadsheet for you to check before using the information. To use it, you simply drop in pictures of receipts, and the page reads the vendor, date, total, and tax information from the text. It highlights any fields it is unsure about, and you can fix those fields before checking the "I checked this" box. Once you've reviewed everything, you can download the receipt data as a CSV file.
You don't need a terminal, account, or API key to use this tool. It's built entirely within your browser, so your pictures are never uploaded anywhere. The project was developed for the DEV's Hacktoberfest 2026 Side Hustle challenge and was built using AI coding assistance. The code was tested and evaluated, with measurements taken on 100 Indonesian receipt photos (CORD-v2 test split, CC-BY-4.0, used for testing only).
The web page uses open reading models (PP-OCRv5 mobile detection and recognition models, Apache-2.0) through PaddleOCR.js and ONNX Runtime Web, which run directly in the browser page. A rules-based step picks out the vendor, date, total, and tax from the text lines. Any unsure fields are marked with "Please check" along with a brief reason. Tax is always marked, as it tends to be the weakest field.
When you download the CSV, it opens in Excel, Google Sheets, or LibreOffice. Text that starts with certain characters, such as =, +, -, or @, is given a leading apostrophe to prevent the spreadsheet from interpreting it as a formula. The reader can handle many receipts at once, running smoothly on a 2 GB test machine. However, it hasn't been tested on phones, and performance might vary on other devices.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.