Line matching makes document changes easy to scan
The comparison finds a longest common sequence of normalized lines, then marks lines outside that sequence as removed from the original or added to the revision. Each displayed line retains its source line number and literal text. The similarity percentage is based on matched lines across both sides, so it is a navigation aid rather than a claim that two documents have the same meaning.
Choose normalization rules deliberately
Exact mode treats capitalization and every whitespace character as meaningful. Optional settings can ignore letter case, trim line edges, collapse whitespace runs, or omit blank lines while matching. The displayed rows always use the original text; normalization only determines whether two lines count as equal. Compare in exact mode first when formatting changes matter.
Local limits prevent the browser from freezing
Each side is limited to one million characters and 2,000 lines, and the comparison matrix is capped at four million line pairs. Imported text files are read locally and limited to 2 MB. The downloadable report uses familiar space, minus, and plus prefixes but is a readable summary, not a patch guaranteed to apply with developer tools.
Text is processed in the current browser and is not intentionally uploaded to Gypes. Still, avoid pasting secrets into any device or browser profile you do not trust, and review consequential legal, code, policy, or contract changes manually rather than relying only on a line match.