JPG to XLSX read a compressed receipt photo into worksheet rows, but nonbreaking-space entities appeared in some text cells.
From real runs
| Element | Result | What happens |
|---|---|---|
| Photo skew | Changed | The slightly rotated receipt was read into an upright worksheet, so the paper angle did not carry into Excel. |
| Column headers | Kept | Item, Qty, and Amount appeared as three cells in the header row of the generated worksheet. |
| Numeric cells | Kept | Quantities and amount values were written as numbers, including normalized negative values for returns and coupons. |
| Receipt spacing | Changed | Some spaces in the register and card-auth rows were serialized as ' ' text in the workbook. |
| Rows without quantities | Changed | Tax and coupon lines kept amount values but used blank cells in the Qty column, producing ragged rows. |
| Background and logo | Lost | The receipt paper color and store-logo-like header styling did not become workbook formatting. |
| Runtime | Kept | The JPEG receipt generated a 16.5 KB XLSX in 8.4 seconds and consumed one Deep Read credit. |
The slightly rotated receipt was read into an upright worksheet, so the paper angle did not carry into Excel.
Item, Qty, and Amount appeared as three cells in the header row of the generated worksheet.
Quantities and amount values were written as numbers, including normalized negative values for returns and coupons.
Some spaces in the register and card-auth rows were serialized as ' ' text in the workbook.
Tax and coupon lines kept amount values but used blank cells in the Qty column, producing ragged rows.
The receipt paper color and store-logo-like header styling did not become workbook formatting.
The JPEG receipt generated a 16.5 KB XLSX in 8.4 seconds and consumed one Deep Read credit.
We converted our JPEG sample (97 KB) with AI Deep Read under AI Deep Read on 2026-09-11. It took 5.6 s and produced a 17 KB Excel XLSX file, 83% smaller than the input. The engine raised 1 warning, listed below.
Input · sample.jpg · 97 KB
1240 × 1754 px, baseline
Output · sample.xlsx · 17 KB
Workbook with 1 sheet (Extracted); first sheet 13 rows, 23 cells, 0 formulas, header filter, fitted column widths
JPG table OCR charges 1 Deep Read credit for each uploaded photo. The converter decodes the JPEG and asks for tabular Markdown before writing XLSX, but compression and spacing can leak into text cells as cleanup work.
Tesseract + pandas 2.3
tesseract receipt.jpg stdout --psm 6 > receipt.txt
python3 - <<'PY'
import pandas as pd
rows = [['Item','Qty','Amount'], ['USB-C cable 2m',1,12.49]]
pd.DataFrame(rows[1:], columns=rows[0]).to_excel('receipt.xlsx', index=False)
PYYou must correct OCR text into rows yourself; this command chain does not solve photographed-table detection.
The HTTP API and the MCP server run the exact engine this page uses (AI Deep Read); the response carries the same warnings the web converter shows. Check support first with GET /api/v1/pairs?from=jpg&to=xlsx.
curl -X POST https://filetypeconverters.com/api/v1/convert \
-H "Authorization: Bearer $FTC_API_KEY" \
-F "file=@input.jpg" -F "from=jpg" -F "to=xlsx" \
-F "deepRead=true" \
--output output.xlsxThe tested JPG workbook serialized extra visual spacing from the OCR transcript as literal nonbreaking-space entities. Replace those strings before importing the sheet downstream.
Yes in the tested receipt. Parenthesized amounts were stored as negative numbers, so '(3.00)' became -3 in the workbook.
The slight rotation in the sample did not stop table detection. Severe perspective, glare, or folded paper can still move values into the wrong cells.