Home / Lab / Text tools

Text tools

Unicode, emoji, RTL, and combining-character counting

How does the current word counter treat emoji sequences, composed and combining accents, non-breaking spaces, zero-width spaces, and text without ordinary spaces?

Written by:

Last tested:

Last reviewed:

Report an error or suggest a correction

Recorded output for Unicode, emoji, RTL, and combining-character counting
Recorded first-party output associated with this report.

Test environment

Chrome 154.0.8037.97 and Edge 154.0.4258.53; 1100 × 850 viewport.

Fixture and expected result

Counts follow the documented whitespace and JavaScript UTF-16 rules, even where those rules differ from grapheme clusters or linguistic words.

Reproduction steps

  1. Download the text fixture and recorded cases.
  2. Paste one line at a time into Word Counter.
  3. Compare words, characters, characters without spaces, sentences, and paragraphs.
  4. Use the receiving platform’s rule when a submission has a strict limit.

Observed result

Both browser runs matched all 12 fixed cases. A family emoji was one whitespace-delimited word and 11 UTF-16 code units; café used four units while the visually similar combining form used five.

Downloadable evidence

The fixture library publishes SHA-256 hashes for exact-byte verification.

What this test does not establish

This is not language-aware tokenization or grapheme-cluster counting. RTL display, screen-reader pronunciation, and platform-specific limits need separate testing.

Related tool and guide

Word & Character Counter · How Long Should a Voiceover Script Be?

Change history

  • : Structured data, image dimensions, and fixture links reviewed.
  • : Report organized under Anvil Tools Lab; evidence and limitations reviewed.