How Text Diff works
A line-based longest-common-subsequence algorithm identifies stable context and classifies additions and removals without sending text anywhere.
The comparison is deterministic: the same files and settings produce the same classification. File contents remain inside the current browser session.