Start a text comparison with the strictest useful settings. First find what changed; only then decide whether differences in whitespace or letter case can be ignored. A quieter result is not necessarily a more accurate review.
Keep two clearly labeled originals
The companion text diff checker compares plain text by lines. Put the earlier version on the left and the later version on the right, or use the corresponding source labels in the interface. Keep both originals outside the tool so that copy-and-paste mistakes can be corrected.
When working from Word, a PDF, or a formatted webpage, copied text can omit comments, formatting, links, or reading order. The diff does not inspect those document features. Export comparable plain text deliberately, and use a document-specific review workflow when formatting or tracked changes are part of the decision.
Run an exact comparison first
- Paste the earlier version into Before and the revised version into After.
- Leave case and whitespace differences visible for the first pass.
- Choose Compare texts and read each removed and added block in context.
- If differences are dominated by formatting, make a second pass using only the ignore options that fit the task.
- Record which settings you used. If you share the result, keep the underlying texts so another reviewer can repeat it.
Earlier: Keep the backup for 30 days.
Later: Keep the backup for 90 days.A line-based diff may show one removed line and one added line. That does not mean two independent editorial changes occurred; it is how the changed line is represented.
Download the synthetic Before text and After text to repeat this example.
The tool normalizes CRLF, CR, and LF line endings even with both ignore options off. A final newline counts as an extra empty line. Inputs are limited to 100,000 characters and 1,500 lines per side. The download is a readable comparison report, not an executable patch or proof of byte-for-byte equality.
Likewise, a paragraph that moves may appear as a removal and an insertion. The tool is not deciding whether the move is intentional or whether two passages mean the same thing.
Use ignore rules only when they match the job
Ignoring whitespace trims each line and collapses its internal whitespace. It can help with extra spacing, but it does not remove or merge line boundaries, so rewrapping paragraphs can still produce differences. It can also conceal meaningful differences in indentation-sensitive code, fixed-width data, or an exact-text requirement. Ignoring case can hide a changed product code or a case-sensitive token. Do not use either setting merely to make a result look cleaner.
Unicode adds another wrinkle: visually similar text can have different underlying character sequences. Unicode normalization is a separate transformation. The presence of a whitespace option does not imply that the tool performs normalization, removes invisible controls, or compares human meaning.
A JSON formatting change is still a text change
These two records may express the same object members in a workflow that does not care about member order, while producing different text:
{"name":"Noa","active":true}
{"active":true,"name":"Noa"}RFC 8259 distinguishes JSON objects from ordered arrays. A plain line diff does not perform schema validation or a semantic JSON comparison. Use the existing JSON formatter to make structure easier to read, while retaining the original copies and remembering that formatting itself changes the text. The formatter preserves numeric and string spellings, including decimal scale and negative zero. Receiving applications may impose their own numeric limits; keep the originals and compare the results.
Review the result, not just the totals
- A tiny edit marks a large block: check whether one input uses long unbroken lines or has different line wrapping. Line boundaries determine the granularity.
- The texts look equal but still differ: inspect punctuation, spaces, Unicode representation, and final line endings. Do not invent a cause from appearance alone.
- The input exceeds the displayed budget: compare smaller meaningful sections with context. Keep section names and boundaries so that omissions are visible.
- No differences remain after ignore options: the result only establishes equality under those selected rules. Run the strict comparison again before an exact-text handoff.
The companion tool compares input locally and does not send it to a processing service. A copied report can still contain private text. Review it before sharing, and treat browser extensions, clipboard history, downloaded files, and host-site analytics as separate considerations. The tool is not a plagiarism detector, legal review, or proof that a document is safe to sign.
