A revision is easier to review when every change has context. Follow the example, then adapt the settings to your own work. Open Text Compare.
Add the original and the revision
Open Text Compare. Under Original text, choose Write or paste text and enter the earlier announcement. Put the revised announcement in Updated text. The original says the workshop starts on Monday and asks attendees to bring a pencil; the update changes the day to Thursday, changes the pencil to a laptop and adds a line about coffee.
You can instead use Upload text file or drop a file onto each card. Imports accept UTF-8 TXT, CSV, JSON, XML, HTML and Markdown, up to 1 MB per source. These formats are read as text. Swap texts reverses the original and updated roles, so use it only when those roles are wrong.
Use the exact example input
ORIGINAL The workshop starts on Monday. Bring a notebook and a pencil. UPDATED The workshop starts on Thursday. Bring a notebook and a laptop. Coffee is included.
For an accurate review, use complete corresponding sections. Missing context can make an addition look like a replacement.
Choose how to match the versions
For this prose example, leave Compare by set to Words and keep both ignore checkboxes off. Select Compare texts. The tool compares words, spacing and punctuation; it does not judge whether the new wording is better.
Use Lines for a list or text-based code, and Characters to inspect small changes. Ignore letter case can hide capitalization-only changes. Ignore extra whitespace treats a whitespace run within a comparison unit as one space, but Lines still uses line boundaries. These choices affect the match and similarity score, so record them when sharing a report.
The example in this guide
Verified with i2TEXT on November 11, 2026, using Words with both ignore settings off.
- Similarity
- 79.2%
- Change groups
- 2
- Added / removed words
- +5 / -2
Read the changes rather than just the percentage
The real downloaded example reports 79.2% similarity, two change groups, five added words and two removed words. The day change is one group. The pencil replacement and adjacent coffee addition form the second group; a change group is not necessarily one edited word.
Side by side keeps both versions visible with red removals and green additions. Unified combines the diff into one flow. Changes only focuses on the changed groups. Use the previous and next controls to move between them.
Similarity is twice the matched comparison units divided by the total units in both versions. Because whitespace and punctuation are units too, it is not simply the percentage of unchanged words. It also does not measure meaning, correctness or plagiarism.
Download the report that suits your review
Select HTML report under Report and use the download icon. The example saves as i2text-comparison.html. It contains the statistics, settings, highlighted sources and a change register. Open this file in a browser to review it independently of the tool.
Choose Text report for a plain-text record or JSON data for structured settings, segments and statistics. Both complete sources are included. The print icon opens a report for printing; choose Save as PDF in the browser dialog if needed. Copy summary copies the measurements rather than producing a merged final draft.
The comparison runs locally in your browser. Save the report before refreshing or closing the page. For a very large or substantially different section, read any broad-comparison notice and split the sources into smaller sections if you need finer detail.
Original and updated announcementsi2text-comparison.html
Common questions
Why are there two groups when I made three edits?
Adjacent additions and removals are collected into a group. In this example, the laptop replacement and new coffee sentence are adjacent in the diff and appear together.
Can I ignore whitespace without changing my files?
Yes. The checkbox changes matching, while the original source spelling and spacing remain in the report. It does not rewrite either source.
Does the report contain the full input?
Yes, within the supported input limits. The on-screen preview can be shortened and a difficult section can be grouped broadly, but the report retains both complete source texts.
Try Text Compare with your own task.
Use your own source and review the result before sharing.