Invisible Character Detector
Paste text to see every invisible or unusual Unicode character with its code point, line and column, the surrounding words, and whether it looks like an artifact or a legitimate part of the language.
Nothing is uploaded · No credits used · Works offline once loaded
Text with hidden characters made visible
Reading the results
Which hidden characters matter, and why
| Character | Typical source | What the detector does |
|---|---|---|
| U+200B Zero Width Space | Web pages, chat apps, copied tables | Marks it safe to remove, except in Thai, Khmer, Lao, Myanmar and Japanese text, where it marks line-break points. |
| U+200C Zero Width Non-Joiner | Persian, Kurdish and Indic spelling | Keeps it between letters of joining scripts; flags it inside Latin words. |
| U+202E Right-to-Left Override | Spoofed file names, “Trojan Source” code | Marks it high severity when the text has no right-to-left script. |
| U+E0000–E007F Tag characters | Hidden prompts aimed at AI tools | Decodes and shows the hidden ASCII text; keeps valid flag emoji such as Scotland and Wales. |
| U+00A0 No-Break Space | Word processors, HTML | Keeps it between a number and its unit and before French punctuation; otherwise suggests a normal space. |
| U+FE00–FE0F Variation selectors | Emoji presentation, hidden data | Keeps one after an emoji; flags long runs, which can encode hidden data. |
Each row in the results lists the code point, how often it appears, a recommendation for every occurrence and the surrounding words, so you can decide in context. The “made visible” view shows the full text with each hidden character replaced by a labelled marker.
The detector also lists words that mix Latin letters with Cyrillic or Greek look-alikes, such as a Cyrillic “а” inside an English word. That trick is used to dodge filters and fake brand names.
When you are ready to remove what you found, open the text in the Unicode Text Cleaner with one click.
Frequently asked questions
What are invisible characters?
Unicode characters that take no visible space or look like ordinary spaces, such as zero-width spaces, joiners, bidirectional controls, tag characters and variation selectors. They change how text behaves when it is searched, counted, sorted or displayed.
Why does my text contain them?
Most come from copy and paste: word processors, PDFs, web pages, chat tools and AI assistants. Some are required by a language, for example the zero-width non-joiner in Persian, and some are added on purpose to hide information.
Can invisible characters be dangerous?
Yes, in some places. Bidirectional overrides can make code or file names display differently from how they run, and Unicode tag characters can carry hidden instructions for AI systems. The detector marks these as high severity and decodes hidden tag text.
Does my text leave my browser?
No. The detector runs entirely in your browser and does not send or store the text.
Does finding invisible characters prove text was written by AI?
No. The same characters come from editors, languages and copy and paste. They are a text-quality signal, not evidence of authorship.