How the receipt scanner works
The receipt scanner reads your photo with optical character recognition (OCR) running entirely in the browser. When you upload an image, Receipt Caker loads a WebAssembly OCR engine, passes it the picture, and gets back the text it can make out — the business name at the top, the date, the line items with their prices, and the totals. That text is then run through the same parser the PDF importer uses, which sorts the lines into a structured receipt: vendor, date, items, tax rate. The result loads straight into the live builder as an editable draft.
Because everything happens on your device, there is no upload and no server round-trip for your image. The trade-off is that OCR reads pixels, not a clean text layer, so the accuracy of the draft tracks the quality of the photo. Receipt Caker leans into that reality by always presenting the scan for review: you check the fields, correct anything the read got wrong, and only then export.
Getting the best scan
On a phone you do not need to save a photo first — tap "Take a photo" and your camera opens straight away, so you can shoot the receipt in front of you and have it read the moment you snap it. On a computer, drag in or upload a photo or screenshot you already have.
A little care with the photo dramatically cuts the correcting you do afterwards. Lay the receipt flat and shoot straight down on it, so the lines are not skewed. Get even light with no harsh shadow across the print, and fill the frame with the receipt rather than shooting it small in a busy background. Sharp focus matters most of all — a slightly blurry photo turns 8s into 3s and drops whole words.
Thermal receipts fade over time, and a faint print is hard for any OCR to read; scanning sooner, while the ink is dark, gives a much better result. If a scan comes back thin, it is usually the image, not the tool — retaking the photo with better light and focus is the fastest fix. Whatever the scan recovers, the editor lets you complete the rest by hand, so a poor photo still saves you from starting on a blank receipt.
Who scans receipts
Anyone who ends up with paper receipts and needs them as data benefits from scanning. Freelancers and sole traders photograph a receipt the moment they get it, so the details are captured before the slip is lost or fades. Small-business owners turn a shoebox of receipts into digital records for their bookkeeper. Employees scan receipts from a trip straight into an expense report. In each case the point is the same: the numbers already exist on paper, and typing them again is wasted effort.
By reading the photo and laying the result out as an editable receipt, the scanner removes most of the keystrokes while keeping you in charge of accuracy. It pairs naturally with the receipts-to-CSV tool for batching many receipts into a spreadsheet, and with the expense report maker for bundling them into a claim.
Private, on-device receipt scanning
Receipts carry personal detail — card last-four digits, addresses, what you bought and where — so the scanner keeps the image on your device. The OCR engine runs locally in your browser, and your photo is never uploaded, stored, or logged on a server. The only thing fetched from the network is the OCR engine itself, the first time you scan; your receipt picture stays put. That keeps sensitive information private and makes the tool fast once the engine has loaded. Since you review every extracted field before exporting, the finished receipt contains only what you have checked and corrected.