Key takeaways
- A replacement contributes one removal and one addition.
- Ignore options affect token comparison while preserving displayed text.
- Each side accepts up to 40,000 characters and 2,000 comparison tokens.
The text difference checker compares an original version with a changed version and marks additions and removals. Use line mode for paragraphs, lists or records arranged one per line. Switch to word mode to inspect smaller wording edits. Everything runs in the current browser without uploading or saving the compared text.
Before comparing, decide which version is the reference. Put that version in Original text and the revision in Changed text. Reversing the inputs reverses the meaning of added and removed, even though the changed content is the same. Start with the ignore options disabled so the first result reflects the text as pasted.
Choose what you want to inspect
Line mode treats each line as a comparison token. A changed word can therefore cause the whole old line to appear as removed and the whole revised line as added. This is useful when each line represents a separate record or sentence, because the surrounding unchanged lines make the edit easy to locate.
Word mode uses non-whitespace runs with their following whitespace as tokens. It is useful for a small revision inside a paragraph. It is not a language-aware word segmenter: punctuation is part of the adjacent run, and languages without spaces may produce long tokens. A punctuation change can therefore count as a token replacement.
Both modes use a longest common subsequence comparison. The tool keeps an ordered set of matching tokens and marks the remaining ones as additions or removals. Repeated identical lines can admit several equally short alignments, so the selected position of a repeated line is not evidence that a particular physical line moved.
Step by step with CrizaTools
- Paste the baseline into Original text and the revision into Changed text. The initial sample can be replaced directly, or use Clear both to start with empty fields.
- Select Line or Word under Compare by. The comparison updates when inputs or options change.
- Select Ignore case if letter casing should not count as a difference.
- Select Ignore whitespace if differences in leading, trailing or repeated whitespace inside comparison tokens should be normalized.
- Choose Side by side or Inline under View. Read the status counts, then inspect the marked result.
- If the tool reports a size limit, compare a smaller corresponding section from each version. Keep a few unchanged lines around the suspected edit for context.
Side-by-side output keeps Original and Changed columns visible. Removed entries have a minus prefix and a strike-through; added entries have a plus prefix and underline. Inline output interleaves the operations with bracketed markers. These textual and structural cues supplement the background colours.
Worked example: one replaced line
Put the following in Original text:
alpha
beta
Put the following in Changed text:
alpha
gamma
Choose Line. The first line matches, while beta is removed and gamma is added. The status reports 1 added, 1 removed lines. It does not call this two independently changed source lines or a single insertion; it describes the operations required to replace the old content.
Switch between Side by side and Inline. The operation counts stay the same because View changes presentation only. If you swap the two inputs, gamma becomes the removal and beta becomes the addition. This example’s operation counts are verified in the logic tests.
Worked example: an insertion and ignored formatting
Compare red blue with red green blue in Word mode. The added token is green , including its following space. Red and blue remain in sequence, so the result reports 1 added, 0 removed word tokens. This demonstrates an insertion rather than a replacement.
For a second check in the same workflow, replace the inputs with Hello world and hello world. Enable both Ignore case and Ignore whitespace. The matching key for the first token becomes hello on each side, and the second token is world on both sides. The result reports no additions or removals.
The side-by-side view still retains the original capitalization and spacing in each column. Ignoring formatting affects equality checks; it does not rewrite the inputs or produce a normalized file. These word-mode cases are also covered by Node tests so the expected counts match the implementation.
Understanding the ignore options
Ignore case converts comparison keys to lowercase. It does not perform translation or decide that two differently spelled terms have the same meaning. Use it for case-only edits such as heading capitalization, and turn it off when case carries meaning in identifiers or another receiving system.
Ignore whitespace trims the token and collapses whitespace within it to a single space for comparison. In line mode, line boundaries remain boundaries. Adding an empty line is still a line operation; the option does not merge a whole document into one stream. In word mode, tokenization happens before normalization, which also matters for standalone whitespace tokens.
Line mode normalizes CRLF and CR line endings into LF boundaries. Consequently, changing only the newline encoding does not create line removals and additions. This tool is intended for visible text comparison and does not claim to identify every byte-level change in a file.
Limits and common mistakes
Each input is capped at 40,000 characters and 2,000 line or word tokens. The bounded comparison uses a table with at most roughly four million cells. There is no large-document background queue or automatic upload fallback. If either cap is exceeded, the status asks you to compare a smaller section instead of attempting an unrestricted calculation.
Blank input is allowed. Comparing empty Original text with a non-empty revision represents additions; comparing a non-empty original with an empty revision represents removals. A trailing line break can create an empty final line token, which may explain an otherwise unexpected extra operation.
Common mistakes include using line mode for a single long paragraph, expecting ignored whitespace to remove blank lines, interpreting repeated-line alignment as move detection and assuming a low edit count proves equivalent meaning. Review the actual words. The tool does not parse programming languages, merge files, generate a patch or explain the effect of a changed instruction.
Questions and answers
Why does one wording change show an addition and a removal?
The old token must be removed and the revised token added. Counts describe edit operations, so a replacement contributes to both totals.
Can I upload text files?
This checker uses pasted text fields. Copy the relevant sections into the two inputs and stay within the character and token caps.
Does switching views change the comparison?
No. Side by side and Inline display the same operations. Compare by and the ignore checkboxes determine how those operations are calculated.
Why are repeated lines aligned unexpectedly?
Several equally short token alignments may exist. The algorithm chooses one ordered alignment and does not identify moved blocks as a separate operation.
Does the checker remember my pasted text?
The tool does not store it locally or send it to a server. Keep your own source copies if you need to compare the same versions later.
Inspect revisions with the text difference checker, or use the word counter when you need document length instead of an edit comparison.
Free tools for this
PDF & image
Text Difference Checker
Compare two texts by line or word, see added and removed counts, and switch between inline and side-by-side views.
PDF & image
Word Counter
Count words, characters with and without spaces, sentences and paragraphs as you type, with reading and speaking time.
PDF & image
Case Converter
Convert text to UPPERCASE, lowercase, Sentence case, Title Case, camelCase, snake_case, kebab-case and more in one click, then copy.