Know which scan pages need another look / KitForma

Check scanned PDFs for missing page numbers and order

A successful scan can still hide a skipped face, a repeated page or the wrong back-page order. Read the printed numbers in your chosen top or bottom band, compare the evidence with your reference and leave with a focused rescan list. Everything runs locally.

Open the workbench Understand the details

No account neededRuns in your browserFree to use

Local evidence / clear next steps01 → 02

Printed sequencePages 1 · 2 · 4Selected by you

Review the scanInspect the gapEvidence before a decision

Your originals stay unchanged. Keep a report of what the check found.

From question to next step

Scan Coverage

Start with the details you know. Unknown answers stay unknown. Try the example to see how the check works.

Your inputs

Required fields are marked with an asterisk.

Scan workflow

Scanned / processed PDF (front pages for duplex)

Local PDF, at most 80 MiB per file and 256 pages total per run. No password-protected files.

Back-page PDF

Keep your original back batch unchanged. Select the ordering below.

Back-batch order

Untouched original PDF

Pages are compared using bounded rendered thumbnails. This is not an exact-content comparison.

Printed-number band

Band height (% of page)

Start with the bottom 12%. Increase only if the printed number is outside the crop.

Horizontal number location

Number-region width (% of page)

Use 100% for alternating left/right folios; narrow the region if it contains dates or several numbers.

Printed numbering style

Use local OCR when no text-layer number is found

An English OCR model is loaded from this site. Your page crops stay on this device. Uncertain readings remain unresolved.

Expected retained PDF pages (optional)

Use an independent count of retained faces, including intentionally blank pages where required.

Expected physical sheets (optional)

Assumes one retained front and back for every sheet. Do not use if blank backs were deliberately removed.

Expected first printed label (optional)

Supply both range ends, for example 1 and 120 or i and xii. Use one section per range.

Expected last printed label (optional)

Intentional numbering resets (optional)

One output-page=first-label per line, e.g. 12=1. A Roman-to-Arabic change is detected automatically. Leave the expected range blank when using resets.

Verified label corrections (optional)

One output-page=label per line, e.g. 4=12 or 5=iv. Use 4=? for an unread page. Output indices refer to the interleaved order in duplex mode.

Loading the local tool… Your files stay in this browser.

Your report

A practical interpretation of your evidence.

Make the next step clear.

Your findings, assumptions and next steps will appear here after you run the check.

Scan Coverage · https://scancoverage.kitforma.com/ · KitForma

Read real PDF text and scanned number crops on your device

Review page-number gaps, duplex counts and original/processed visual candidates

Download the full page ledger and a focused CSV review queue

Go one level deeper

Understand the evidence.

Focused guides explain the rules, the edge cases and the checks worth doing before you act.

Field guide / 01

Find missing page numbers without inventing missing pages

Use printed folios, independent counts and a page-level ledger to focus your scan review.

Read the guide ↗Field guide / 02

Check front and back scan batches before interleaving

Compare face counts and test reversed back order without changing your original PDFs.

Read the guide ↗Field guide / 03

Audit page loss after blank removal and scan cleanup

Compare original and processed PDFs with counts, local rendered evidence and a review queue.

Read the guide ↗

Good questions

Before you start

Does my PDF leave this device?

No. PDF parsing, page rendering, OCR and visual comparisons run locally. The PDF engine and English OCR assets load from this site. Reports you download contain filenames, page labels and file hashes, so share them deliberately.

Can it prove that no pages are missing?

No. A source count, expected printed-label range or untouched original helps identify discrepancies. Numbered pages can still contain missing content, and a never-scanned unnumbered page cannot be inferred from the remaining PDF. The report never certifies content completeness.

Does it work with image-only scans?

Yes. When the selected region has no usable PDF text-layer label, the tool renders that number crop and runs local OCR. A result below the confidence threshold, several competing numbers or a failed page remain unresolved. Review the crop and correct the label yourself when needed.

What do I do with Roman numerals or chapter resets?

Roman-to-Arabic changes create separate numbering sections. For a reset within the same style, enter the output page and its first label under section starts. An expected first/last range applies to a single section; do not combine it with several resets.

Will it change the order or delete blank pages?

No. The duplex mode constructs a review order and labels each face in the report. It does not write or rearrange your PDF. A blank-looking page can be intentional or contain faint information, so this tool does not remove it.

How reliable is the original-to-processed comparison?

It matches low-resolution rendered page evidence, not PDF bytes or full-resolution content. A close result is a visual candidate. Similar forms, blank pages, crops, rotations and edits can cause ambiguity; inspect those originals manually. The file-level SHA-256 identifies the exact PDF used for the report.

The evidence behind the check

Sources & assumptions

Rules are reviewed against the references below. Manufacturer updates, firmware and your exact setup can change the result.

PDF.js API reference ↗
Local byte input, page rendering, text content and resource configuration.
Tesseract.js API documentation ↗
Local OCR workers, recognition output and worker lifecycle. OCR confidence is not a correctness certificate.
NAPS2 profile settings ↗
Blank-page thresholds can remove nonblank pages; keep an original batch and review exclusions.
University of Michigan digitization specifications ↗
An archival workflow reference for preserving page order and documenting missing originals; not a certification performed by this tool.

Reviewed 2026-10-13. Read our methodology or report an issue.