Using this tool takes only a few seconds. Paste the original version of your text into the left panel labeled "Original Text" and paste the updated version into the right panel labeled "New Text." Click the Compare button to run the comparison. The tool will immediately display statistics — how many lines were added, removed, and unchanged — along with a similarity percentage. Three display modes are available via the tabs above the output: Side by Side shows both versions in parallel columns with changes highlighted in their respective panels; Unified shows a single column using + for additions and - for removals, similar to a git diff; Summary shows a larger statistical overview in card format. If you prefer the diff to run continuously while you edit, check the "Compare automatically as I type" box. Use the Copy Diff button to copy the full unified diff with + and - prefixes to your clipboard, or use Copy New Text to grab only the final version of your text.
A text diff is a representation of the differences between two versions of text. The term "diff" comes from the Unix command-line utility of the same name, first released in 1974, which compared files line by line and output a minimal description of how to transform one into the other. Today, diff is embedded in nearly every developer tool and document management system — from Git and GitHub to Google Docs' version history and Microsoft Word's Track Changes feature. Under the hood, most diff algorithms are built on the Longest Common Subsequence (LCS) principle: they find the largest set of lines that appear in both texts in the same order, then treat everything outside that common set as either an addition or a deletion. This approach minimizes the apparent number of changes, making the diff as readable as possible for humans.
Writers use text diff tools to track changes between drafts of articles, scripts, or manuscripts — especially when collaborating with editors who return a revised copy without explicit change tracking enabled. Legal professionals use them to compare contract revisions, quickly spotting clauses that have been added, removed, or subtly reworded between negotiation rounds. Developers reach for diff tools when reviewing configuration file changes, comparing API responses across environments, or checking rendered HTML output when refactoring templates. Content managers use text comparison when updating web pages, to verify that only the intended sections were changed during a CMS migration or editorial update. Students use diff to compare their submitted essay draft with instructor feedback, and teachers use it to check whether submitted work was heavily revised from a previous version.
The Longest Common Subsequence algorithm works by building a two-dimensional matrix where each cell stores the length of the LCS for all elements up to that point in both sequences. For two texts of m and n lines respectively, the matrix has (m+1) × (n+1) cells. Each cell is filled by looking at whether the corresponding lines match: if they match, the value is one more than the diagonal cell; if they do not match, the value is the maximum of the cell above and the cell to the left. Once the matrix is complete, the algorithm backtracks from the bottom-right corner to reconstruct the actual sequence of equal, added, and removed lines. Lines that appear in the LCS are marked as unchanged; lines in the original text that are not part of the LCS are deletions; lines in the new text that are not part of the LCS are additions. This is the same approach used by Git's diff engine, making the output familiar to anyone who has used version control.
For writers, a diff tool offers a lightweight alternative to word processor track changes when collaborating across different applications or platforms. If you send a draft in plain text or Markdown and receive a revised file back, pasting both versions here instantly shows exactly what changed. You can verify that your editor only touched the sections you expected, confirm that an approved paragraph was not accidentally deleted during reformatting, or compare a before-and-after version from a client who did not use change tracking. The similarity percentage gives you a quick read on how heavily a version was revised — a score above 90% indicates light edits, while anything below 70% suggests significant restructuring. Checking word-level changes within modified lines can reveal subtle substitutions, such as a tonal shift from "should" to "must" in a policy document.
Developers often need to compare text outside of a version control context. Common scenarios include comparing two environment configuration files to confirm a staging environment matches production, diffing API JSON responses from two different versions of an endpoint to verify backward compatibility, checking rendered HTML output before and after a template refactor, comparing log snippets from two separate incidents to find recurring patterns, or reviewing database migration scripts from two branches before merging. Pasting both versions here is often faster than setting up a terminal diff command, especially for quick one-off checks during debugging. The unified diff output produced by the Copy Diff button is in standard format and can be pasted directly into documentation, bug reports, or code review comments.
A few small adjustments can make your diff output significantly more useful. First, normalize line endings before comparing — if one text uses Windows-style CRLF and the other uses Unix LF, every single line will appear as changed even though the content is identical. Most text editors have an option to convert line endings. Second, decide on the right level of granularity for your use case: comparing complete paragraphs as single lines will show you which paragraphs changed but not what within them changed, while comparing at the sentence or word level gives more precision but more noise. Third, if you want a clean diff for copy-pasting into documentation, use the Unified view and click Copy Diff — this produces output in the standard + / - format readable by humans and tools alike. Fourth, if two texts are more than 80% different, consider whether you are actually comparing the right pair of documents; a very low similarity score often means the files are mismatched rather than heavily edited.
Expert guides and educational resources to help you master the math.
Lorem ipsum comes from Cicero's 45 BC philosophical text — deliberately scrambled so reviewers focus on layout, not words. Here's the full history, how it became a design standard, and when to stop using it.
