Syrian users read and type both sets of digits, and even the names are tangled: what English calls Arabic numerals, Arabic usually calls something else. Here is what the term means and how an app should treat both.
Arabic-Indic digits are the digit shapes ٠ ١ ٢ ٣ ٤ ٥ ٦ ٧ ٨ ٩, Unicode U+0660 to U+0669. They have the same values as 0 to 9 but are different characters, so software must convert them before calculating or searching. Chrome and Node use them by default for Syrian Arabic (ar-SY).
The names are where people trip. In English, "Arabic numerals" means 0 to 9, and Unicode calls ٠ to ٩ "Arabic-Indic". Arabic often turns this around: ٠ to ٩ are «الأرقام الهندية», the Indian digits, and 0 to 9 are «الأرقام العربية». Both names point at the same history: the digits began in India and reached Europe through Arabic, and each side named them after the people it learned them from. To keep things clear, this page says Arabic-Indic for ٠ to ٩ and Western digits for 0 to 9.
There is a second set. Extended Arabic-Indic digits, U+06F0 to U+06F9 (۰ ۱ ۲ ۳ ۴ ۵ ۶ ۷ ۸ ۹), are what Persian keyboard layouts type. Most look like the first set and a few, such as ۴ and ۶, are drawn differently, but all ten are separate code points. English names for the two sets overlap ("Eastern Arabic numerals" is used for either), so code points are the precise way to talk about them.
| Name | Notes | |
|---|---|---|
| English | Arabic-Indic digits | Unicode's name. Also "Eastern Arabic numerals" and, from the Arabic, "Hindi numerals". "Arabic numerals" usually means 0 to 9. |
| Modern Standard Arabic | الأرقام الهندية | The usual formal name for ٠ to ٩, with 0 to 9 called «الأرقام العربية». Some writers swap the two to match English. |
| Also heard in Syria | «أرقام عربي» and «أرقام إنكليزي» | Everyday speech often calls ٠ to ٩ the Arabic digits and 0 to 9 the English ones, the reverse of the formal names. |
| In code or standards | U+0660 to U+0669, U+06F0 to U+06F9 | Arabic-Indic and Extended Arabic-Indic digits. A locale tag asks for them with -u-nu-arab, or for Western digits with -u-nu-latn. |
A classifieds app in Latakia asks sellers for a price and a contact number. One seller types the price ١٢٥٠ and the number ٠٩٣٣١٢٣٤٥٦. Another types 1250 and 0933123456. A third, on a Persian keyboard layout, types ۱۲۵۰. To the app these are different strings, yet they mean the same thing.
All of them should end up as the same stored values: the phone number in Western digits and the price as an integer. Our tested parser turns ١٬٢٥٠٫٥٠ (with the Arabic thousands and decimal separators), 1,250.50 and 1250.5 into the same 125050 hundredths, and ۱۲۵۰ into 125000 (tested 16 November 2026). A description that mentions ١٢٥٠ is found by a search for 1250, because the search copy is converted to Western digits too, and the screen can still show the price as ١٬٢٥٠٫٥٠ ل.س.
| Digits | What we saw | |
|---|---|---|
| ar-SY in Chrome or Node | Arabic-Indic | ١٬٢٥٠ for 1250 |
| ar-SY-u-nu-arab | Arabic-Indic | ١٬٢٥٠٫٥٠ ل.س. as currency |
| Plain ar in Chrome or Node | Western | 1,250 for 1250 |
| ar-SY-u-nu-latn | Western | 1,250.50 ل.س. as currency |
| Flutter, intl 0.20.2, locale ar | Western | 1,250.50 with ل.س. as currency |
0x0660, each Extended one its code point minus 0x06F0, which is exactly what our tested helpers do.-u-nu-arab or -u-nu-latn on the web, a fixed format in Flutter), and keep the same choice for prices, dates and counters.One display bug belongs to these digits alone. In our Chrome 157 test on 16 November 2026, the phone number ٠٩٣٣ ١٢٣ ٤٥٦, typed with spaces between the groups and placed in a span with dir="ltr", was drawn with its groups in reverse order. The reader sees the last group first and may copy the wrong number.
The cause is the bidirectional algorithm, which classes ٠ to ٩ as Arabic numbers and lets them pull the spaces between groups toward right-to-left, even inside a left-to-right element. Two fixes passed in the same test: unicode-bidi: isolate-override on the element, or writing the number as one run with no spaces.
The digit shapes ٠ ١ ٢ ٣ ٤ ٥ ٦ ٧ ٨ ٩ (Unicode U+0660 to U+0669). They mean the same as 0 to 9 but are different characters, so apps convert them before doing anything with the number.
Digit output checked with Node 24 (ICU 78), Chrome 157 and flutter test (intl 0.20.2) on 16 November 2026.
Tell us where it happens: a form, a receipt, an SMS. We build Arabic-first apps and will help you find the cause.