Regex to Extract Emoji from Text — Unicode Property Escapes

Extracting Emoji from Text

Emoji detection looks deceptively simple but requires Unicode property escapes to handle multi-codepoint sequences correctly. Modern JavaScript (ES2018+) supports \p{Extended_Pictographic} with the u flag.

Single Emoji

\p{Extended_Pictographic}

(Use the u flag.) Matches single-codepoint emoji like 😀 and 🎉.

Including Skin-Tone Modifiers and ZWJ Sequences

Emoji like 👩‍💻 (woman technologist) are sequences joined by zero-width joiners (\u200D). To capture them as one match:

\p{Extended_Pictographic}(?:\u200D\p{Extended_Pictographic})*

For full coverage including emoji modifiers (skin tones), variation selectors, and regional indicators (flags):

(?:\p{Extended_Pictographic}(?:\uFE0F|\u20E3)?\p{Emoji_Modifier}?(?:\u200D\p{Extended_Pictographic}(?:\uFE0F|\u20E3)?\p{Emoji_Modifier}?)*)

Tested Examples

Input	Single	Sequence-Aware
"hi 😀"	😀	😀
"👩‍💻 is a developer"	👩, 💻	👩‍💻
"🇯🇵 Japan"	🇯, 🇵	🇯🇵
"thumbs up: 👍🏽"	👍, 🏽	👍🏽

Counting Visible Emoji

JavaScript’s text.length returns UTF-16 units, not characters. Use a sequence-aware regex to count emoji:

const count = [...text.matchAll(
  /\p{Extended_Pictographic}(?:\u200D\p{Extended_Pictographic})*/gu
)].length;

Stripping Emoji

text.replace(/\p{Extended_Pictographic}/gu, "")

Practical Recommendation

For accurate emoji handling in production (segmentation, normalization), use Intl.Segmenter with granularity grapheme. Reserve the regex approach for scanning, not analysis.