Text emoji normalize
Exacttext_emoji_normalizeDeterministic and authoritative — the canonical answer for this input.
Detect, remove, or replace emoji in text. Strategies: 'remove' (drop silently), 'replace' (substitute a configurable placeholder), 'describe' (annotate inline with codepoint). Returns emoji_found sidecar list with char, codepoint, and position — for auditing. ZWJ-joined sequences handled as single clusters. Processing multiple inputs? Use `text_emoji_normalize_batch` to do up to 1000 in one call.
Parameters
textstringrequiredstrategystringoptional'remove', 'replace', or 'describe'.
replacementstringoptionalUsed when strategy='replace'.
Call it
curl -X POST https://api.snipget.ai/v1/text/emoji/normalize \
-H "Authorization: Bearer sk_live_..." \
-H "Content-Type: application/json" \
-d '{"text":"Great work team! 🎉🚀","strategy":"remove"}'The Batch tab sends up to 1000 inputs in one call via /v1/text/emoji/normalize/batch.
Example response
{
"input_length": 19,
"output_length": 17,
"text": "Great work team! ",
"emoji_found": [
{
"char": "🎉",
"codepoint": "U+1F389",
"position": 17
},
{
"char": "🚀",
"codepoint": "U+1F680",
"position": 18
}
],
"emoji_count": 2,
"strategy": "remove"
}FAQ
Why strip or normalize emoji?
Emoji can break search, sorting, and parsers that expect plain text, and they often sneak into names or addresses pasted from chat. This cleans them out, or swaps them for a placeholder, in one pass.
What are the three strategies?
'remove' drops emoji silently, 'replace' swaps each for a placeholder you set with replacement, and 'describe' annotates each emoji inline with its codepoint. It also returns an emoji_found list so you can audit what was there.
Does it handle combined emoji like family or flag sequences?
Yes. Joined sequences (such as a multi-person family emoji) are treated as a single cluster rather than torn into pieces.
Does it change other special characters?
No. It only touches emoji. For curly quotes, accents, or whitespace use the typographic, unicode fold, or whitespace tools.
Is it reliable to run repeatedly?
Yes. It is deterministic and authoritative: the same text and strategy always produce the same cleaned output.