Chunkify Unicode Text

No. The chunker runs entirely in your browser. Text you paste never touches the network – no fetch, no XHR, no analytics on content.

Graphemes are what humans see: 👨‍👩‍👧 is one grapheme. Code points are what Unicode defines: the same sequence is five code points. UTF-16 units are what JavaScript stores internally: the same sequence is eight UTF-16 units. Different questions, different answers – the tool lets you pick.

In graphemes mode (default), yes – including skin-tone modifiers (👨🏾), ZWJ families (👨‍👩‍👧), and flags (🇬🇷). In code-points mode, single emoji stay intact but ZWJ families and flags can split. In UTF-16 mode, even a single 😀 can break into two halves.

It’s the only mode that matches a human reader’s intuition. The tool uses Intl.Segmenter with granularity:'grapheme', which the Unicode standard defines for exactly this use case.

Code-points mode when you need a fixed-size-per-character count and the text has no ZWJ sequences. UTF-16 mode when you are interoperating with APIs that count string.length (like Twitter’s old 140-UTF-16-unit limit) and want to match their view.

The tool returns an error – you cannot split N units into more than N chunks. The error names the unit type (graphemes / code points / UTF-16 units) so you can check your assumption.

Yes. Whitespace is counted like any other unit and never stripped. Newlines inside the input are treated as normal characters and appear inside whatever chunk they fall into.

Yes – all Unicode text works: Latin, CJK (Chinese/Japanese/Korean), Arabic, Hebrew, Devanagari, Greek, Cyrillic, and every other writing system.

No hard limit. The tool handles 10,000-character inputs in well under 30 ms on a modern laptop. Very long inputs in graphemes mode stay fast because Intl.Segmenter is native C++ code in the browser.

After the page loads, yes. HTML, CSS, and JS are self-contained – disconnect Wi-Fi and keep chunking.