Loading page
About this tool
Enterprise-grade, privacy-first document version comparison & verification tool. Compare PDF, DOCX, TXT, JSON, CSV, HTML, XML & Markdown locally with metadata analysis, tampering detection, visual diffs, and AI insights.
The Document Version Verifier compares two versions of the same document in the browser: it extracts the text layer from PDF, DOCX, TXT, JSON, CSV, HTML, XML or Markdown, runs a line-level diff between them, and independently fingerprints each file with CRC32, MD5, SHA-1, SHA-256 and SHA-512 so a byte-identical pair is proved by a matching SHA-256 rather than by eye. Alongside the diff it scores tampering cues out of 100 — author or creating-software changes, a created date later than the modified date, a drop in page count, zero-width hidden characters, embedded scripts — and reports Low, Medium, High or Critical risk. Files are read locally and never uploaded.
Open Document Version Verifier on AltFTool — it loads instantly in your browser.
Press Select Files, or drop files on 'Drop your documents here or click to browse' — up to 10 files, 200MB max per file, accepting .pdf, .docx, .txt, .rtf, .csv, .json, .xml, .html and .md; Try Demo loads two sample agreement versions instead.
Assign '1. Original Document (Version A)' and '2. Updated Document (Version B)' from the uploaded list — the comparison pipeline reruns whenever that selection changes, with no submit button.
Work through the tabs: Content Comparison for the line diff, Hash Verification for the CRC32, MD5, SHA-1, SHA-256 and SHA-512 digests, Security for the tamper score, then Export PDF, JSON, CSV, TXT or HTML from the Export tab (Export PDF saves document-version-audit-report.pdf).
CRC32, MD5, SHA-1, SHA-256 and SHA-512 are all computed locally, so you can match whatever checksum a counterparty already published while relying on SHA-256 for the actual verdict.
Each signal adds a weighted amount to a capped 100-point tamper score — for instance 35 points when version B claims a creation date before version A — which rolls up into a Low to Critical risk level.
Line-ending, trailing-whitespace and whitespace-collapsing options are optional and disclosed with the result, so you always know which differences the diff was told to ignore.
Compare their SHA-256 hashes: if the two digests match, the files are byte-for-byte identical and the tamper score is 0 with risk level None. Matching file sizes or a diff showing no visible changes is weaker evidence, because formatting and metadata differences can sit outside the text layer.
PDF, DOCX, DOC, RTF, TXT, Markdown, JSON, CSV, HTML and XML, up to 200 MB per file, with PDF text extraction bounded to the first 100 pages. Each comparison representation is capped at 250,000 characters and 3,000 lines.
An author mismatch adds 25 points, a change of creating software 15, a reduced page count 20, zero-width or hidden characters 25, embedded scripts or macros 30, and an impossible timestamp 30 to 35. A total of 60 or more is reported as Critical, 35 to 59 High, and 15 to 34 Medium.
No. The score flags cues that deserve a human look — many have innocent explanations, such as a file re-saved in different software or converted between formats. Treat it as a prompt to check provenance, and for anything legally or financially significant get the document examined properly.