NORTHMELD WORKFLOWS
Turn scanned PDF tables into Excel data
Scanned PDFs store page images rather than usable table text. Northmeld can use enhanced cloud recognition for pages that need OCR, then lets you review the extracted tables before exporting Excel or CSV.
For teams receiving image-based reports or scanned statements that need editable table data.
How it works
- 1
Choose a readable scan
Use a complete, upright PDF with legible text. Open it in the workspace and choose enhanced recognition when needed.
- 2
Inspect OCR results
Check extracted cells against the page image. Pay particular attention to decimal points, dates, identifiers, and similar-looking characters.
- 3
Confirm and export
Correct uncertain values and table structure, then export the reviewed rows to XLSX or CSV.
Example input and output
Input
Scanned page: Part code | Amount 00128 | 1,250.00
Output
Part code,Amount 00128,1250.00
Illustrative target structure. Preserve identifier leading zeros and verify number formats in your spreadsheet application.
Limits and review
- Blurred text, handwriting, skewed scans, and dense layouts can reduce recognition quality.
- OCR is not guaranteed to be exact; review the output against the original.
- Enhanced recognition uses private cloud processing and may consume cloud points. See current plans for availability.
Common questions
How is a scanned PDF different from a text PDF?
A scanned page is an image. OCR is needed to turn its characters into machine-readable text; a text PDF already contains usable text.
Is OCR always applied to every page?
Northmeld reads embedded text first and escalates pages that need OCR, rather than treating every page as a scan.
Can I rely on the result without checking it?
No. Confirm important numbers, dates, and identifiers against the source before using or exporting the result.