Skip to content
Coverton Blog
In this article
Documents and text

Normalize Unicode in text that looks identical

Accented characters or Korean text can look identical while using different underlying character sequences. Normalization can align their representation, but names and meaningful symbols still need a careful comparison with the source.

Open the toolHow to use this tool ↓
Unicode normalize
This illustration explains the tool's workflow. It is not the processed result of an uploaded file.

How to use this tool

  1. Enter text or add the source text file.
  2. Select the required normalization form.
  3. Run and compare names and special symbols.
  4. Download the normalized text.

Choose the right settings

Choose NFC or NFD according to the character composition needed by your workflow. NFKC and NFKD also change compatibility characters, so review full-width forms and special symbols before applying either to sensitive identifiers.

Supported input
  • TXT,MD
Available settings
  • Normalization form

Before you start

Choose content or file organization

Decide whether the task concerns wording, document structure, hidden information or filenames. Keep an original copy and check the supported input. Extracted plain text is useful for reviewing words, but it does not reproduce every picture, font, equation or page arrangement in the source.

Where separators, normalization or text cleanup are available, try a short passage with headings, unusual characters and meaningful line breaks. For inspection or file comparison, include all relevant source files instead of assuming the visible page tells the whole story. For renaming copies, confirm the selected order and keep the mapping between old and new names.

Before text cleanup, decide which spaces, repeated lines or characters carry meaning; a spelling difference may be intentional. Where the task is file inspection or organization, record the files included and the literal naming or comparison choices instead. To diagnose a problem, compare a short source passage or a known file pair with the result. Check an intermediate copy separately: a previous extraction, normalization or rename may explain the difference before this tool is involved.

Check meaning as well as appearance

Read the result’s beginning, ending and transitions between sections. Compare names, table values and special characters with the source. If the tool produces a report, locate sample findings in the original file; an inspection result is a review aid, not automatic removal of the identified information.

For downloadable file copies, verify names, extensions and correspondence to the originals before using them elsewhere. If wording is missing or jumbled, distinguish unsupported embedded content from an encoding or ordering issue. A filename change does not change the underlying format, and a matching appearance does not prove two files have identical bytes.

Example input and output

decomposed text + NFC → composed text

Examples describe the workflow. Results depend on the source file and settings.

Check the saved result before using it

Inspect lines containing Korean, accents, full-width characters and symbols alongside their originals. When the visual result seems unchanged, test searching or string comparison in the program that will use the normalized text.

If the result is different from what you expected

Normalization cannot repair characters already damaged by incorrect file decoding or a corrupt source. If the text is visibly broken, investigate the original file and its encoding before changing Unicode representation.

Limits to keep in mind

  • NFC is suitable for normal composed text. NFKC/NFKD also change compatibility characters. Review the result.

Continue with a related tool

Extract text from DOCX

Extract DOCX body text and supported headers or footers when you need to review wording. The TXT result does not reproduce fonts or page layout, so compare its content separately from the document's printed appearance.

Choose blank lines, single-line separation or section markers for the extracted parts. Those markers describe extracted sections and do not recreate the DOCX's physical printed page numbers.

Open the tool

Extract text from PPTX

A TXT extract is useful for reviewing the wording of a presentation slide by slide. Do not assume the extracted sequence matches visual placement on the canvas; inspect how titles and explanations connect in the output.

Choose the slide-marker mode to add a SLIDE label to each extracted section. Use blank-line or single-line separation when those labels are unnecessary for the text you will edit.

Open the tool
CovertonUnicode normalize
Open the tool

This guide explains the public browser tool. Keep a copy of the original and inspect the result before sharing it.