How do you track receipts without an app? Skip another expense download. The app-free system is not a Camera Roll album or a Drive folder named Receipts. It is a sheet with date, vendor, total, and tax on every row, plus a link back to the slip. Soft product land: something that fills those columns so you are not retyping forever.
Tax week. Phone has coffee snaps with no useful filename. Inbox has twelve vendor PDFs. Drive has Receipts 2026 full of files that open and close and still are not a ledger. Folders feel free of apps. They store files. They do not give you a sum you can filter.
The honest fork
A blank Google Sheet or Excel workbook is fine if you will type every line. Date, vendor, total, tax, source. That is a real ledger with zero new apps. The pile is the problem: paper in the bag, PDFs in email, snaps in Photos. Typing scales until it does not. An expense-app gallery that never exports rows you own is the other trap. Neither folder nor gallery is the sheet.
| date | vendor | total | tax | source |
|---|---|---|---|---|
| 2026-04-03 | Amazon Web Services | 428.17 | 0.00 | 01-aws.png |
| 2026-04-18 | Blue Bottle Coffee | 18.47 | 1.52 | 04-bluebottle.png |
| 2026-04-27 | Staples | 67.34 | 5.54 | 06-staples.png |
| 2026-05-12 | Home Depot | 156.88 | 12.91 | 10-homedepot.png |
Copy that shape. One row equals one slip. Keep tax its own column when the slip shows it. Link the PDF or photo. Do not paste a card-statement total as if it were the merchant receipt.
Why a Drive OCR dump still fails
Searching a folder of OCR text feels clever. Grepping still leaves a human to pick which dollar amount is the ledger total. A local Q2 2026 run on n=20 synthetic receipts (Tesseract 5.5.0 via pytesseract, grayscale + 2× upscale, dumb OCR / Drive-like extract) shows why columns beat a text dump. Vendor hit 95% (19/20). Date and total hit 100%. Tax hit 60% (12/20). All four fields correct on only 55% (11/20). Ground-truth Q2 sum $2,034.43 matched the OCR-parsed sum (absolute error $0). Treating OCR like a searchable folder, every receipt (20/20) still had 2+ currency-like amounts, avg 6.05 matches per receipt (121 matches across the corpus). Raw OCR text is not a receipt ledger.
What “done” looks like without another expense app
Capture slips when they arrive: photo, PDF, or forward into one place. Land date, vendor, total, tax, source. Review odd tax blanks. Filter and sum when someone asks for March or Meals. Export CSV if you need a copy outside the workbook. OCR fills digits. You still decide the row.
Slipsheet turns photos and PDFs into rows you review, then a sheet or CSV you own. Photo, PDF, or forward to receipts@slipsheet.app. Soft land only: not another approval suite. The product is the columns. The folder was never the system.
A few straight answers
Is a Camera Roll album enough to track receipts without an app?
No. It stores photos. Tracking means structured rows you can filter, sum, and hand over without opening every image.
Can I stay in Google Sheets or Excel only?
Yes, if you type or import into date, vendor, total, and tax. The workbook is the app-free ledger. The Drive folder of PDFs is not.
Does OCR in Drive replace the spreadsheet?
No. OCR text often has several dollar amounts per slip. Someone still picks the total and tax. Columns are the ledger.