Invisible characters
Give hidden marks a name before deciding what to change.
Hidden is not the same as unwanted
Text can contain characters that do not draw a visible symbol. Some represent unusual spacing. Others control joining, presentation, or direction. Removing every such character can change the text rather than simply clean it.
Use Inspect characters to see named categories and context. The app's targeted removal rules are off initially. Meaningful joiners, variation selectors, emoji components, and direction controls remain intact in launch transformations.
Characters you can choose to change
| Character | What selection does | Consequence to review |
|---|---|---|
| U+00A0 non-breaking space | Convert to an ordinary space. | Changes whether text can break at that gap. |
| U+200B zero-width space | Remove it. | May join words or change break opportunities. |
| U+00AD soft hyphen | Remove it. | Discards discretionary hyphenation intent. |
| U+FEFF BOM / no-break character | Remove it from entered text. | File-signature handling during import is separate. |
The app does not silently include all similar characters in a removal rule. A narrow non-breaking space, word joiner, and ordinary space are distinct characters.
Why emoji need joiners and selectors
The emoji 👩💻 includes U+200D ZERO WIDTH JOINER between components. Family emoji can also be multi-character sequences. Variation selectors can request a text or emoji presentation. The Unicode emoji specification describes these intentional sequences.
Removing a joiner can make a combined emoji display as separate components. Removing a selector can change its presentation. Text Cleaner retains both rather than treating them as accidental clutter.
Reference: Unicode Emoji, including presentation selectors and ZWJ sequences.
Languages and direction marks
U+200C ZERO WIDTH NON-JOINER can carry language-specific joining intent. Combining accents can be separate code points attached to a preceding letter. Direction controls can affect how mixed-direction text is presented. These are not automatically deleted.
The inspector reports selected direction controls without interpreting whether each is needed. Text direction in the workspace changes display only; it does not insert or remove directional characters.
What inspection can and cannot tell you
The table lists the code point, name, count, cleanup status, and first context of each reported character. A marker such as ⟦U+200B⟧ is a display aid in the inspector, not text inserted into the result.
Inspection covers a defined collection, not every invisible character or every suspicious string. It does not certify safety for HTML, SQL, terminal commands, or source code. The app treats entered markup as text; downstream applications have their own rules.
Choose a targeted repair
- Open the character inspector and identify the actual code point.
- Keep the original and look at the context where it occurs.
- Select only the applicable removal or conversion rule.
- Review Cleaned or Compare for joined words, changed spacing, and emoji presentation.
Read the inspector instructions. If your problem is an ordinary gap, start with the space-cleanup guide.