Question 1

When should I compare by word instead of by line?

Accepted Answer

Line mode is right when the text has meaningful line structure: code, config files, CSVs, anything you would review as a diff. Prose usually arrives as a handful of very long lines, so any edit marks a whole paragraph as changed. Word mode ignores line boundaries entirely and marks just the words that moved, which is what you want when comparing two drafts of the same paragraph. Character mode is for short strings, keys and IDs, where a single transposed digit is the thing you are hunting for.

Question 2

Why is one line shown as edited but another as removed and re-added?

Accepted Answer

A removed line and an added line are paired up and highlighted word by word only when they still share at least 30% of their text. Below that they are two different lines that happen to sit in the same place, and highlighting the handful of characters they have in common would obscure the change rather than show it. That is also why swapping a line for a completely new one shows as a clean remove plus add.

Question 3

What do the ignore options actually compare?

Accepted Answer

"Ignore case" lowercases both sides before matching, so Title Case and sentence case count as the same line. "Ignore whitespace" trims each line and treats any run of spaces or tabs as one space, which is what you want after a reformat or a re-indent. Both affect matching only: the panels always show your text exactly as you pasted it, with its own capitalisation and spacing intact.

Question 4

Do Windows and Mac line endings show up as changes?

Accepted Answer

No. CRLF, LF and lone CR are all read as a line break, so a file saved on Windows compares cleanly against the same file saved on macOS or Linux. A trailing newline at the end of a file is ignored too, since it is a save-time convention rather than an edit.

Question 5

Can I compare two files instead of pasting?

Accepted Answer

Yes. Drop a file on either box or use the Upload button on that side. Plain formats work directly (.txt, .md, .csv, .tsv, .log, .json, .xml, .srt, .vtt) and .docx files have their text extracted with paragraph breaks preserved, so two revisions of the same document line up. PDFs and old .doc files can't be read, so export those as .docx or plain text first.

Question 6

What does the Copy button give me?

Accepted Answer

A unified patch: the whole comparison as one block of text with removed lines prefixed "-", added lines prefixed "+", and unchanged lines indented. That is the format code review tools and email threads expect, so it stays readable when the colours are gone.

Question 7

How do I find the changes in a long file?

Accepted Answer

Long stretches of untouched text collapse to a "Show N unchanged lines" bar, leaving three lines of context on each side of every change, which is the same amount a patch file carries. Click a bar to open that stretch on its own without giving up collapsing everywhere else. The arrows above the result step from one change to the next and wrap around at the end, and the counter tells you which one of how many you are looking at.

Question 8

What does the similarity percentage mean?

Accepted Answer

The share of the comparison that came through unchanged, measured in whatever unit you are comparing by. Two drafts of the same article typically sit in the nineties in word mode; anything under about 50% usually means you are comparing two different texts rather than two versions of one.

tooldura

Text Diff Checker

About Text Diff Checker

How to use Text Diff Checker

Frequently Asked Questions