What is U+202F?
U+202F is the narrow no-break space (NNBSP). It is a fixed-width space, roughly half as wide as a normal space, that tells a text renderer never to wrap a line at that spot. It is not invisible like a zero-width character. It is a real, visible gap that simply looks the same as a normal space to the eye, which is why it causes so much confusion.
| Character | Code point | Width | Line break allowed? |
|---|---|---|---|
| Space | U+0020 | Normal | Yes |
| No-break space | U+00A0 | Normal | No |
| Narrow no-break space | U+202F | About half | No |
Where does U+202F come from?
- ChatGPT, April 2025. Days after OpenAI released o3 and o4-mini, the startup Rumi reported those models emitting U+202F where ordinary spaces belong. OpenAI described it as a quirk of large-scale reinforcement learning rather than a watermark. It stopped within days. Our ChatGPT hidden characters guide covers the wider picture.
- French and other typeset text. French typography uses a narrow no-break space before ; : ! ? and around guillemets, so French web pages, PDFs and Word files carry it on purpose.
- Locale-formatted times and numbers. Unicode CLDR release 42 (2022) began using U+202F between a time and AM or PM in English formats. Software built on the newer locale data can therefore output a time whose space is not U+0020, which broke string comparisons in some projects. Check your own runtime rather than assuming.
- Copy and paste from formatted sources. Word, Google Docs and web pages can carry it across when you copy. See why text pastes weird.
What does it break?
Anything that compares or searches for a normal space will miss it: exact-match searches, spreadsheet lookups, string equality in code, URL slugs and file names. It also survives "trim" functions in many languages, because it is not the ASCII space they look for. If two strings look identical but do not match, suspect U+202F or the ordinary non-breaking space (U+00A0).
How do I find and remove it?
Paste your text into the cleaner above. It reports each U+202F it finds and swaps it for a normal space. To do it yourself:
| Where | How |
|---|---|
| JavaScript | text.replace(/\u202F/g, " ") |
| Python | text.replace("\u202F", " ") |
| Excel | =SUBSTITUTE(A1, UNICHAR(8239), " ") |
| grep (GNU, PCRE) | grep -P "\x{202F}" file.txt |
For the full set of look-alike gaps, see the invisible Unicode characters list and the general weird spacing fix.
When should I keep it?
Keep U+202F in text that will be typeset as French, or where a number and its unit must never split across lines. Replace it when the text is going somewhere that treats it as data: code, spreadsheets, search, databases, or file names. In CleanPastedText, turn off "Exotic spaces" in Advanced Controls to leave it alone.