Diff PDF, Word, TXT and Markdown files side by side. Word-level highlighting, CJK-aware tokenisation so Chinese text is compared by meaning rather than by byte, plus both a text view and a layout view, with exportable reports.
Every operation runs in your browser using WebAssembly and Web Workers. Your file is never uploaded to a server, so nothing can leak, be logged, or be used to train anything.
Contracts and translated documents rarely differ line-by-line. This compares at word level and handles Chinese segmentation, so a single changed character in a clause is found rather than marking the whole paragraph as rewritten.
Yes — both are parsed to normalised text first, so cross-format comparison works.
No. Both documents are parsed and compared inside the browser tab.