Text Compare
Compare two texts or files to identify differences, similarities, and changes with detailed analysis
About the Text Compare Tool
A diff shows exactly what changed between two pieces of text, rather than leaving you to spot the difference by eye. This is the same idea used by version control systems, applied to any two blocks of text you paste in.
You can compare line by line for documents and configuration files, or at word and character level when the changes are small and fine-grained.
How to use it
- 1Paste the original text into the first box and the revised version into the second, or upload two files.
- 2Choose the comparison mode: line level for code and config, word or character level for prose and small edits.
- 3Review the highlighted output — additions and deletions are marked, and matching lines are shown side by side for context.
- 4Check the similarity percentage for a quick sense of how much actually changed, then copy the diff for your notes.
When you would use it
- Reviewing exactly what changed between two versions of a config file or a contract clause.
- Checking that a translation or a rewrite kept every original sentence.
- Confirming that nothing was dropped when text was copied between two documents.
How it works, and what it does not do
The comparison runs on our server: the input is split into lines or words, aligned with a diff algorithm, and returned as a sequence of unchanged, added and removed runs alongside a similarity percentage. Because it works on the text you submitted and nothing else, the differences it reports are real — it cannot be fooled by formatting that is not present in the input.
The similarity figure is a rough guide, not a judgement about meaning. Two paragraphs that say the same thing in different words will score low, while two that share most of their wording but invert a requirement — “must” becoming “must not” — will score very high. Read the diff itself; treat the percentage as a sorting hint.
Frequently asked questions
Which comparison mode should I use?
Use line mode for source code, JSON, YAML and configuration files, where each line is a meaningful unit. Use word or character mode for prose, where a single changed word inside a paragraph would otherwise mark the entire line as different.
How is the similarity percentage calculated?
It measures how much of the two texts matches after alignment. A high percentage with a small diff means a light edit; a low percentage means the texts have diverged substantially, and a line-level diff will be easier to read than a character-level one.
Why does my diff show everything as changed?
Almost always a whitespace difference. Changing indentation from spaces to tabs, or adding a trailing newline, makes every line differ. Normalising the formatting before comparing usually collapses the diff to the real changes.
Is the text I paste stored anywhere?
No. Both texts are sent to our server for comparison, held in memory for the duration of the request, and discarded. Nothing is written to disk or logged with its contents.
Something is flagged as different but the lines look identical.
The differences you cannot see are almost always invisible ones: trailing whitespace, a leading tab that became spaces, or line endings that differ between the two files. Copying each side through the tool again from the original source usually resolves it.
How do the line and word modes differ?
Line mode reports whole lines as added or removed, which suits code and configuration. Word and character modes find the smallest change inside a line, which is what you want when comparing prose that has been lightly edited.