One visible symbol does not always equal one unit in a character counter. Some emoji use multiple Unicode code points, and JavaScript strings store text as UTF-16 code units. Those are different ways of measuring the same displayed text.
The ordinary letter A has a JavaScript string length of one. The emoji 😀 has a length of two. TheDailyUtils character total uses JavaScript string length, so those two inputs will not receive the same count even though each appears as a single visible symbol.
A family emoji or an emoji with a skin-tone modifier can contain several components joined into one visual presentation. A counter based on visible grapheme clusters can therefore disagree with one based on code units. MDN explains the distinction in String.length.
Paste the final text into the word counter, then into the actual form. If the counts differ, use the form's own rule for that submission. Replacing an elaborate emoji with a simpler symbol or plain words may reduce the count, but verify rather than guessing.
Do not remove combining characters indiscriminately. An accent can be part of a person's name, and changing it just to satisfy a counter can change the text. Keep an untouched copy, shorten ordinary wording first, and review the result in its destination. A count is a measurement of text, not a judgment about how many symbols a reader perceives.