Truncate, Pad and Wrap: Shaping Text to a Fixed Width

After reading this you will know how to cut text to a character limit without breaking words, pad lines so they align in a monospace column, and hard-wrap paragraphs at a fixed width, and you will understand why counting characters is trickier than it looks.

What the tool does

Three operations share one idea: they make text conform to a fixed width. Truncating cuts a long string down to a maximum length. Padding adds filler characters so every line reaches the same length. Wrapping breaks a long paragraph into lines no wider than a set number of columns.

Here is one hook. Take the line The quick brown fox jumps over the lazy dog, which is 43 characters. Truncate it to 20 characters at a word boundary and add an ellipsis, and you get The quick brown…, which is 16 characters: three whole words plus the single ellipsis character. Notice that the result is shorter than 20, not exactly 20. Word-boundary truncation gives up length to avoid cutting brown in half. That trade-off is the whole story of this tool.

When to use it, and when not

Reach for truncation when a display has hard limits: a table cell that shows 40 characters, a preview snippet, a database column capped at 255. Reach for padding when you are building fixed-width output by hand: aligning a list of names, faking a monospace table, formatting a punch-card-style record. Reach for wrapping when you paste a paragraph that arrives as one giant line and you need it broken at column 72 for an email or a commit message.

Do not use hard-wrapping for text that a screen will re-flow on its own. A web page, a word processor and a chat app already wrap soft. If you insert real line breaks at column 72 and the reader's window is narrower, the text wraps twice and looks ragged. Hard-wrap only when the destination uses a fixed width and will not wrap for you.

Counting characters: code points, not bytes

Every operation here depends on a length, and length is ambiguous. A "character" can mean a byte, a Unicode code point, or a grapheme (what a human sees as one symbol). These three disagree.

Take the string café. If the é is stored as the single code point U+00E9, the string is 4 code points but 5 bytes in UTF-8, because é takes two bytes. If instead the é is stored as e plus a combining accent (U+0065 U+0301), the string is 5 code points and still displays as 4 symbols. An emoji like 😀 is one code point but 4 bytes in UTF-8.

This tool counts Unicode code points. That matches what most programming languages report for string length and it treats one emoji as one character. It does not count grapheme clusters, so a combining accent or a flag emoji built from two regional-indicator symbols may count as more than one. The Unicode Standard defines these levels in Annex #29 on text segmentation, which is the reference to check when the counts surprise you.

The rules behind each operation

Truncation to a limit L keeps the first L code points and discards the rest. With an ellipsis, the ellipsis itself must fit inside the budget, so the kept text shrinks:

Here k is the number of code points kept from the original, L is your limit, and \text{len}(\text{ellipsis}) is the length of the marker you append. The single character … (U+2026) has length 1, so truncating to 20 with an ellipsis keeps 19 original characters plus the marker, reaching 20 total. If you use three dots ... instead, that marker has length 3 and you keep only 17.

With the word-boundary option on, the cut moves left to the last space at or before position k, so the result can be shorter than L. Left padding to width W adds W - n filler characters in front of a line of length n; right padding adds them after; centre padding splits them, putting the extra one on the right when W - n is odd.

Wrapping walks a paragraph word by word and starts a new line whenever adding the next word plus one space would exceed the column limit C. A word longer than C is left intact on its own line rather than split, so URLs and long tokens survive.

The chart below shows how the kept length depends on your limit and whether the word-boundary option is on, for this exact sentence with a one-character ellipsis.

Reading the results

Three things are worth checking after any run. First, the actual length of each output line. Padding guarantees a minimum width, never a maximum: a line already longer than W is left untouched, not cut. If you need both a floor and a ceiling, truncate first, then pad. Second, whether the word-boundary retreat cost you more than you expected. A limit of 20 that lands one character into a long word can drop that whole word, leaving you well short. Third, whether wrapping preserved your paragraph breaks.

Common mistakes

The most frequent error is counting the ellipsis twice. If you want the final string to be at most 20 characters, the tool already subtracts the marker length for you; do not also shorten your limit by hand. A second mistake is padding with a multi-character filler and expecting an exact width. Padding ab to width 5 with the two-character filler xy cannot land on 5 exactly by repetition; the tool fills to the target and trims the last repeat, so you get xyxab (5), not xyxyab (6).

A third mistake is hard-wrapping text that already contains hard breaks, then wrapping again. Each pass treats existing line breaks as paragraph boundaries, so re-wrapping narrower does not undo the first wrap. If you need to change an existing wrap width, join the lines back into paragraphs first. A fourth: assuming truncation trims trailing spaces. It does not. If your kept text ends in a space, that space counts toward the length and stays before the ellipsis.

Try it: watch the limit reshape the result

Related tools

Shaping width is one step in a cleanup pipeline. To change letter case before or after shaping, use the Case Converter. To search and replace patterns, including collapsing runs of spaces before you pad, use Find & Replace with Regex. If your text carries accented letters that you want reduced to ASCII before a byte-sensitive truncation, run the Accent & Diacritic Remover first, since that lowers the byte-versus-code-point gap. For flipping or mirroring rather than resizing, see Reverse & Upside-down Text.

Related tools

Case Converter

camelCase, snake_case, Title Case, UPPER, lower and more.

Line Tools: Dedupe, Sort & Shuffle

Remove duplicate lines, sort, reverse, shuffle, number.

Accent & Diacritic Remover

Turn café, naïve and Škoda into cafe, naive, Skoda.

Find & Replace with Regex

Search and replace across text, plain or regex, with preview.

Reverse & Upside-down Text

Reverse letters, words or lines; flip text upside down.

Fancy Text Generator

𝓗𝓪𝓷𝓭𝔀𝓻𝓲𝓽𝓲𝓷𝓰, sᴍᴀʟʟ ᴄᴀᴘs, zalgo and other Unicode styles.