My text diff tool reports thousands of differences but the documents look almost identical — what is going on?

Usually whitespace and line-break differences. Two PDFs of the same content can have radically different underlying text streams if one was exported at one page size and the other at another, because the line breaks fall in different places. A naive line-by-line diff sees every line as changed even when the words match. The fix: use a word-level diff (not line-level), normalise whitespace before comparing (collapse multiple spaces to one, strip trailing whitespace from each line), and ignore line breaks entirely in the diff algorithm. ScoutMyTool Compare PDF uses word-level diff with whitespace normalisation by default. The result reads as "no changes" if the words are the same, regardless of how the lines fall.

How do I compare two versions of a contract for redlining?

Use a text-diff tool that produces a redline output: deletions struck-through in red, additions underlined in blue, unchanged text in normal weight. This is the customary format in legal markup and matches Word's "track changes" rendering. ScoutMyTool Compare PDF produces redline output directly. For complex contracts with formatting changes (font, indent, headings) that text-diff cannot capture, complement with a visual diff so you can verify that the only changes are textual. The visual diff catches stealth modifications (someone changed a definition style or added an indent) that text-diff might miss.

Can a visual diff find subtle layout changes that look identical at first glance?

Yes — that is the point of pixel-level diff. The tool overlays both PDFs at the same scale and highlights any pixel that differs by more than a threshold (typically 1–3 pixels of position difference, 5–10 RGB-value difference). Result: a heatmap showing every region of every page where the documents differ visually. Useful for marketing proofs where someone shifted a logo by 5 pixels, or for design reviews where an image was replaced with a near-identical version. The threshold is tunable: too tight and JPEG compression noise shows as false-positive differences; too loose and real changes get missed. Start with the default and adjust based on results.

Is comparing two signed PDFs safe — will it invalidate the signatures?

Comparing PDFs is a read-only operation; it does not modify either file, so signatures remain valid. What you should not do: extract content from one signed PDF, modify it, and re-save — that creates a new file with no valid signature. Comparison is safe; merging signed PDFs is also safe (signatures remain valid on their original pages); editing signed pages invalidates the signature on those pages.

How do I compare two scanned PDFs of the same paper document?

Two-step workflow. First, OCR both PDFs to get text layers (ScoutMyTool Make PDF Searchable). Second, run a text-diff on the OCR'd output. The text-diff will surface differences in the underlying text — but be aware that OCR introduces noise (95–98% accuracy on clean scans means dozens of false-positive differences in a 100-page document). Two mitigations: re-OCR both PDFs with the same tool and settings so OCR errors are consistent on both sides (they cancel out in the diff); and complement with a visual diff so you can see whether reported text differences correspond to real document changes or to OCR noise.

Can I compare a PDF to a Word document?

Not directly — different formats, different page models. Two paths. First, convert the Word document to PDF and compare PDF-to-PDF; this is exact but only useful if the layouts match. Second, extract text from both (PDF text extraction + Word text extraction to .txt) and compare the text content; this ignores layout but catches content differences. For redline reviews of a Word draft against a PDF reference (a common legal workflow), use the second path. For visual review of a marketing PDF against its Word source, the first path is better.

What is the privacy story when comparing PDFs online?

Most online PDF-compare tools upload both files to their server, process the diff there, and return the result — which means both files transit through a third party. For confidential documents (legal redlines, financial statements, M&A drafts), this is meaningful exposure. ScoutMyTool Compare PDF runs entirely in your browser tab: both files load locally, the diff algorithm runs in JavaScript, the result displays without either file leaving the machine. Always check whether a PDF-compare tool is client-side or server-side before uploading sensitive material; the difference matters more than the diff algorithm specifics.

6 min read

Compare PDFs — find differences between two versions

By ScoutMyTool Editorial Team · Last updated: 2026-05-20

A vendor sent me an "updated" contract last quarter with no track changes — just two PDFs, version 1 and version 2, "let us know if anything looks different". Comparing two PDFs by eye is unreliable; the changes that matter most are often the quietest. A proper PDF diff tool surfaces every word-level, layout-level, and structural difference in seconds. This article walks through the four diff modes, when to use each, and the specific workflow for contract redlining vs design review vs forensic / authenticity verification.

Four diff modes — what each catches

Mode	What it detects	When to use
Text diff	Word-level additions, deletions, and changes in the text content	Contract redlines, draft vs final document review, legal markup
Visual / pixel diff	Any visual change: text, images, layout, colour, even subtle position shifts	Design reviews, print proofs, marketing PDF approvals where layout matters
Structural diff	Changes in PDF page count, page order, form fields, annotations, bookmarks	Audit-trail review, detecting whether a document was structurally modified after signing
Metadata diff	Changes in PDF title, author, producer, modification date, permissions	Forensic / authenticity review; identifying whether a PDF was edited after issuance

Step by step — redline a contract

Open ScoutMyTool Compare PDF at scoutmytool.com/pdf/compare-pdf and drag both files in (older version on the left, newer on the right).
Pick text-diff mode. The tool runs word-level comparison with whitespace normalisation. Result shows deletions struck-through and additions underlined.
Review each change in context. Click each highlighted change to jump to its location in both PDFs. Note any change you accept, dispute, or want to discuss with the counterparty.
Export the redline. Download a "redline PDF" combining both versions with deletions and additions visible, suitable for sharing with the counterparty or your client.
Run a visual diff as a sanity check. Catches layout-level changes (font swap, indent change, image substitution) that text-diff misses. A text-clean redline plus a clean visual diff together means the document changed only as you saw.

PDF for lawyers: redline workflow in the broader document-review context.
PDF redaction guide: pair with compare to verify redactions did not change document content elsewhere.
PDF metadata editor: inspect metadata before forensic comparison.
PDF security guide: confirm signatures remain valid after comparison.
Searchable PDF: OCR step before comparing scanned PDFs.
Share PDFs securely: for sharing the redline back to the counterparty.
ScoutMyTool Compare PDF: the tool — free, client-side, no upload.

FAQ

My text diff tool reports thousands of differences but the documents look almost identical — what is going on?: Usually whitespace and line-break differences. Two PDFs of the same content can have radically different underlying text streams if one was exported at one page size and the other at another, because the line breaks fall in different places. A naive line-by-line diff sees every line as changed even when the words match. The fix: use a word-level diff (not line-level), normalise whitespace before comparing (collapse multiple spaces to one, strip trailing whitespace from each line), and ignore line breaks entirely in the diff algorithm. ScoutMyTool Compare PDF uses word-level diff with whitespace normalisation by default. The result reads as "no changes" if the words are the same, regardless of how the lines fall.
How do I compare two versions of a contract for redlining?: Use a text-diff tool that produces a redline output: deletions struck-through in red, additions underlined in blue, unchanged text in normal weight. This is the customary format in legal markup and matches Word's "track changes" rendering. ScoutMyTool Compare PDF produces redline output directly. For complex contracts with formatting changes (font, indent, headings) that text-diff cannot capture, complement with a visual diff so you can verify that the only changes are textual. The visual diff catches stealth modifications (someone changed a definition style or added an indent) that text-diff might miss.
Can a visual diff find subtle layout changes that look identical at first glance?: Yes — that is the point of pixel-level diff. The tool overlays both PDFs at the same scale and highlights any pixel that differs by more than a threshold (typically 1–3 pixels of position difference, 5–10 RGB-value difference). Result: a heatmap showing every region of every page where the documents differ visually. Useful for marketing proofs where someone shifted a logo by 5 pixels, or for design reviews where an image was replaced with a near-identical version. The threshold is tunable: too tight and JPEG compression noise shows as false-positive differences; too loose and real changes get missed. Start with the default and adjust based on results.
Is comparing two signed PDFs safe — will it invalidate the signatures?: Comparing PDFs is a read-only operation; it does not modify either file, so signatures remain valid. What you should not do: extract content from one signed PDF, modify it, and re-save — that creates a new file with no valid signature. Comparison is safe; merging signed PDFs is also safe (signatures remain valid on their original pages); editing signed pages invalidates the signature on those pages.
How do I compare two scanned PDFs of the same paper document?: Two-step workflow. First, OCR both PDFs to get text layers (ScoutMyTool Make PDF Searchable). Second, run a text-diff on the OCR'd output. The text-diff will surface differences in the underlying text — but be aware that OCR introduces noise (95–98% accuracy on clean scans means dozens of false-positive differences in a 100-page document). Two mitigations: re-OCR both PDFs with the same tool and settings so OCR errors are consistent on both sides (they cancel out in the diff); and complement with a visual diff so you can see whether reported text differences correspond to real document changes or to OCR noise.
Can I compare a PDF to a Word document?: Not directly — different formats, different page models. Two paths. First, convert the Word document to PDF and compare PDF-to-PDF; this is exact but only useful if the layouts match. Second, extract text from both (PDF text extraction + Word text extraction to .txt) and compare the text content; this ignores layout but catches content differences. For redline reviews of a Word draft against a PDF reference (a common legal workflow), use the second path. For visual review of a marketing PDF against its Word source, the first path is better.
What is the privacy story when comparing PDFs online?: Most online PDF-compare tools upload both files to their server, process the diff there, and return the result — which means both files transit through a third party. For confidential documents (legal redlines, financial statements, M&A drafts), this is meaningful exposure. ScoutMyTool Compare PDF runs entirely in your browser tab: both files load locally, the diff algorithm runs in JavaScript, the result displays without either file leaving the machine. Always check whether a PDF-compare tool is client-side or server-side before uploading sensitive material; the difference matters more than the diff algorithm specifics.

Citations

ISO 32000-1:2008 — "Document management — Portable document format" — base PDF specification.
Myers, E. — "An O(ND) Difference Algorithm and Its Variations" — the foundation of most modern text-diff implementations.
Hunt & McIlroy — "An Algorithm for Differential File Comparison" — the original diff algorithm paper.
Adobe Acrobat Pro documentation — "Compare two versions of a PDF" feature reference.
Image SSIM (Structural Similarity Index) — IEEE TIP 2004 — pixel-level visual similarity metric used in visual diff tools.

Compare two PDFs without uploading them

ScoutMyTool Compare PDF runs in the browser. Confidential contracts and draft documents stay on your machine.

Open Compare PDF tool →