DocDiff compares documents entirely within your web browser. When you upload files, they are read into your browser's local memory only. All text extraction, diffing, redline generation, and export operations run locally on your device using JavaScript. No document content is transmitted to any server for these operations.
Specifically: plain text comparison, PDF rendering, change list generation, redline view, sidebar statistics, and all export formats (Word .docx, .txt, clipboard) are fully local and zero-network.
DOCX Paragraph Parsing (Vercel / parse-docx)This parsing service (api/parse-docx) runs as a serverless function on Vercel infrastructure operated exclusively for DocDiff. It uses open-source libraries (mammoth.js + cheerio) to extract paragraph structure from your DOCX files and returns that structure to your browser. No document content is written to disk, stored in a database, or retained after the function returns. Vercel's serverless functions are stateless by design — data exists only for the duration of the function invocation.
This service is what enables accurate multi-page diff highlighting. Without it, DocDiff falls back to a fully local extraction engine (mammoth.js in-browser) — you will see a yellow "Local engine" badge in the header when running in fallback mode.
If you cannot transmit documents outside your environment under any circumstances, use the Text view, which is always 100% local and never contacts any server.
DOCX → PDF Conversion (Gotenberg)This conversion service (Gotenberg) is self-hosted infrastructure operated exclusively for DocDiff. It is not a third-party SaaS product with its own data retention. The service converts your DOCX to PDF entirely in memory and returns the PDF to your browser. No document content is written to disk, stored in a database, or logged at any point during this process. The conversion service discards all data immediately after returning the PDF.
If the conversion service is unavailable, DocDiff automatically falls back to a fully local DOCX renderer (docx-preview) — your document will still display without any server contact.
To opt out of server-assisted rendering entirely, use the Redline or Text views, which are always 100% local.
No Accounts, No Tracking, No AnalyticsDocDiff does not require an account. There is no login, no email capture, and no user profile. DocDiff does not use cookies, does not run any analytics scripts (no Google Analytics, Mixpanel, Segment, or similar), and does not display advertising.
No telemetry or usage data is collected. DocDiff has no knowledge of which documents you compare, how often you use the app, or what results are produced.
Progressive Web App (PWA) & Offline CachingDocDiff is installable as a Progressive Web App. When installed, a service worker caches the application shell (HTML, CSS, JavaScript, and CDN libraries) locally on your device for offline use. This cache contains only application code — never your documents or comparison results. You can clear this cache at any time through your browser's site settings.
Third-Party LibrariesDocDiff loads the following open-source libraries from public CDNs on first use (cached locally by the service worker thereafter): mammoth.js (DOCX text extraction), docx-preview (DOCX rendering), JSZip (ZIP/DOCX parsing), and PDF.js (PDF rendering). These CDN requests do not transmit your documents — they deliver application code only.
Changes to This PolicyIf DocDiff's data handling practices change materially, this policy will be updated and the version date below will reflect the change. The core commitment — that your documents are never stored or used beyond the comparison you initiate — will not change without explicit notice.