Image-only input
There is no text layer to copy. The result depends on what is visible in the scanned page.
SCANNED PDF TO EXCEL
Convert scanned PDF tables to Excel.
Use the free converter for image-only table PDFs. Review every workbook against the source before using it.
A scan needs source-aware review.
OCR can make visible tables editable. It cannot restore a column that was cropped, a value hidden by blur, or information that never reached the page.
Editable output
Use the workbook to speed review, then check its structure and values against the original image.
PUBLIC EVIDENCE PACK
Four scan conditions, one retained failure.
Each condition uses the same synthetic layout. The saved workbooks are actual benchmark outputs from one dated run, not a general accuracy claim.
Clean image-only source
A raster-only table with no selectable text. This saved run retained the five declared headers and three synthetic rows.
- Declared cells
- 20 of 20 declared cells
- Complete rows
- 3 of 3 complete rows
Moderate resample and skew
A controlled resample, light blur, and 1.2 degree skew. It is one documented condition, not a guarantee for every scan.
- Declared cells
- 20 of 20 declared cells
- Complete rows
- 3 of 3 complete rows
Low contrast, blur, and skew
A stronger controlled degradation with 2.4 degree skew. Review placement and values before using any workbook downstream.
- Declared cells
- 20 of 20 declared cells
- Complete rows
- 3 of 3 complete rows
Right-edge crop failure
The source omits Amount USD and three values. A conversion cannot reconstruct content that is absent from the scanned page.
- Declared cells
- 16 of 20 declared cells
- Complete rows
- 0 of 3 complete rows
The crop failure stays visible.
The right-edge source removes Amount USD and the values 425.30, 1180.00, and 690.75. Its saved workbook has 16 of 20 declared cells and zero complete rows. The correct action is to obtain a complete source page.
Review before the workbook moves downstream.
These checks help when a table will inform a financial, operational, or record-keeping workflow.
Keep the source beside the workbook
A scan can look readable while a border, decimal, or column boundary has shifted. Compare the source PDF before using the output.
Check headers, dates, and amounts
The public example has five headers, three synthetic rows, dates, and Amount USD values that can be checked against its declared reference workbook.
Rescan incomplete pages
A crop, unreadable edge, or missing page is an input problem. Obtain a complete source instead of asking OCR to infer missing columns.
What this example does not prove.
It covers four controlled variations of one synthetic layout. It does not establish performance for handwriting, camera perspective, other languages, naturally noisy archives, or semantic field mapping.
The reference workbook is declared expected data. It is not a model output or a customer document. Test representative documents from your own workflow before relying on a result.
Questions before you convert
Can a scanned PDF become an editable workbook?
A scanned table can be converted into an editable workbook when its visible content can be read. Review the rows, columns, and values against the source before relying on it.
What is different about a scanned PDF?
An image-only scan has no selectable text layer. OCR reads the visible page, so blur, skew, faint lines, crops, handwriting, and unusual layouts can affect the result.
Why is the cropped example shown?
It documents a failure boundary. The missing Amount USD column and values are absent from the source, so the saved workbook cannot recover them.
What are the free converter limits?
The public converter accepts one PDF per guest per day, up to 5 MB and 3 pages. The limit is displayed before a file is processed.
Start with a small scan, then compare it.
The public converter accepts one guest PDF per day, up to 5 MB and 3 pages. Download and review the workbook beside the source.
Open free converter