Text Diff Checker: Compare Two Texts by Line, Word or Character

Paste the original on the left and the changed text on the right. Line mode compares whole lines and is the right choice for code, configuration, logs and lists. Word mode marks only the words that changed inside a paragraph, which suits prose, contracts and documentation. Character mode marks single characters, useful for spotting a typo in an identifier or a changed digit. Character mode compares what a reader sees as one character, so an emoji or an accented letter is never cut in half.

The result can be read side by side, with the original and the modified text next to each other, or as one unified column. Removed text is struck through and added text is underlined, and line views carry a minus or plus marker, so the differences do not depend on telling red from green. Long unchanged stretches are folded to a few lines of context; one switch shows everything.

Ignore whitespace treats lines (or, in word and character mode, whitespace) that differ only in spaces, tabs or line breaks as equal, which hides indentation and wrapping noise. Ignore case makes Hello and hello equal. Line endings are normalised, so a file saved with Windows CRLF compares cleanly against one with Unix LF, and a missing final newline is not reported as a change.

The copy button produces a unified patch, the format printed by diff -u and git diff: two header lines and hunks that start with @@ -start,count +start,count @@, followed by lines prefixed with a space, a minus or a plus, with three lines of context. Paste it into a ticket, a code review comment or a chat. When whitespace or case is ignored, the patch describes the lines as they appear in the modified text.

The comparison uses the jsdiff library, an implementation of the Myers O(ND) algorithm from 1986 that finds a shortest set of insertions and deletions. It runs in your browser and nothing is uploaded. Before the algorithm runs, the identical start and end of the two texts are cut off, which makes the common case of a small edit in a big file instant.

Very large inputs are handled in chunks: word and character mode first compare by line and then look inside only the changed blocks. If two huge texts are almost entirely different, the algorithm would take too long, so after a time limit the tool shows the affected block as replaced as a whole and says so. A diff finds a minimal edit, not the edit a human would call a move: moved paragraphs appear as a removal and an addition.