THE SMALL DETAILS, REVEALED

Make the invisible visible.

Find hidden Unicode characters. Understand what they do.
Clean your text without losing what makes it yours.

Zero uploads.Zero guesswork.Inspect first. Clean with care.
LOCAL TEXT INSPECTION
Text stays in this tab
0 / 1,000,000 UTF-16 unitsInspected automatically

See what’s hiding in your text.

Paste something on the left. Invisible characters will appear as labeled markers here.

Review Context mattersLine breaks preserved

A careful cleanup.

Choose what changes. Keep what matters.

Removes only a leading BOM and NUL characters. Preserves zero-width spaces, emoji, joining controls, directional controls, variation selectors, typography, tabs, and line breaks.

Before—code points
After—code points
ChangesWaiting for text

What we found 0 types

Detected does not automatically mean unwanted.

Your findings will appear here, with names, explanations, and exact locations.
A CLEARER LOOK AT YOUR TEXT

Three steps.
Nothing left to guess.

Copied text can carry more than words. A hidden break, a special space, or a direction mark can explain why two identical-looking strings behave differently.

Paste the puzzling text

Bring a sentence, a spreadsheet cell, or a copied paragraph. Inspection runs automatically in a local browser worker.

See each hidden character

Read its name, code point, occurrence count, and locations. Amber markers invite review. Teal markers remind you that context matters.

Keep control of the cleanup

Start with conservative Safe mode, or select specific categories in Advanced mode. Compare the counts, then copy or download the result.

👩‍💻 → 👩‍💻ZWJ preserved in Safe mode
U+1F469ZWJU+1F4BB
INVISIBLE DOESN’T MEAN UNNECESSARY

Some hidden characters
are doing important work.

A zero-width joiner can hold an emoji together. A non-joiner can preserve correct letter shaping. A direction mark can make a mixed Arabic and English sentence readable.

That’s why Safe mode removes only a leading byte order mark and NUL characters. Other detected characters stay intact until you explicitly choose a category in Advanced mode. Review the result whenever typography or language matters.

Read about Unicode emoji sequences ↗
GET TO KNOW THE USUAL SUSPECTS

A field guide to hidden text.

Small characters. Real consequences.
A FEW GOOD QUESTIONS

Before you
clean that text.

The right change depends on the context.

What counts as a hidden character?

This tool checks a curated catalog of zero-width characters, special spaces, joining controls, directional formatting controls, variation selectors, emoji tag letters, script fillers, and non-layout controls. Ordinary spaces, tabs, and line separators are preserved. It does not detect every Unicode property or decide whether text is malicious.

Why does Safe mode leave zero-width spaces alone?

A zero-width space can be an unwanted copy-and-paste artifact, but it can also mark valid word-break opportunities. Safe mode preserves it. To remove it deliberately, switch to Advanced and select Zero-width separators. That category also removes word joiners and BOM characters, so review all findings first.

Will emoji and multilingual text survive?

Safe mode preserves ZWJ, ZWNJ, variation selectors, directional controls, combining marks, and script-specific formatting. Advanced removal can change shaping, spelling, reading order, or emoji appearance. The marker preview interrupts these sequences to reveal their components. The cleaned output retains them unless their category is selected.

Does my text leave this page?

Text is passed only to a local browser worker. The application does not upload, log, or store your input. Reloading clears the editor. Copy and download transfer the result to your clipboard or a local file at your request. Hosting access logs, browser extensions, and your operating system have their own behavior.

How much text can I inspect?

Up to 1,000,000 UTF-16 units per inspection. Detection and cleaning cover the full input. The highlighted preview shows at most 2,500 code points, and location details show the first 12 occurrences per character type. Counts and downloads include the entire input. A worker keeps the inspection off the main thread.

Why are code point counts different from what I see?

One displayed symbol can contain several code points. A family emoji may contain people and joiners, and an accented letter may contain a letter and a combining mark. Supplementary code points use two UTF-16 units. Locations report both code point positions and UTF-16 offsets to make that distinction explicit.