Produce placeholder text in paragraphs, words, sentences or characters, in five different word styles. The text is generated in the page from small built-in word banks, so it works offline and nothing you type is uploaded.
Filler text exists so that a layout can be judged on its structure rather than its words. The trouble with the canonical Latin passage is that it is so recognisable that reviewers stop reading it entirely, which is useful for judging a page and useless for judging whether a headline works or a button label fits. Offering several word styles gives you text with the same visual weight that does not trigger that trained blindness.
A practical sequence for testing a page template with a long body area, a short summary box and a fixed-length metadata field. All three need different output from the same tool.
Every style draws from a fixed list held inside the page. The lists are short — around fifty to sixty words each — and that constraint is visible in the output once it runs past a paragraph or two. Sentences are between eight and twenty words long and paragraphs between five and ten sentences, so a three-paragraph result is well past the point where the bank starts repeating itself. For a single block that is invisible. For a full-page mock-up it is not, and the honest description of the result is Latin-flavoured placeholder rather than convincing prose.
| Alternative | Notes |
|---|---|
| Judging line length and readability | Unreliable. Placeholder words are short and repetitive, so a column of it looks tidier than a column of real English. A measure that holds sixty characters of Latin will feel cramped with technical prose. |
| Judging heading hierarchy and visual weight | Reliable, and the reason to switch styles. Headings always carry real content while body copy does not, so the relationship between the two is visible even when the paragraphs are not being read. |
| Testing character limits and truncation | Reliable in Characters mode, which guarantees the exact length. Unreliable in the other three, where the count varies between runs and a limit may never be reached. |
| Leaving placeholder text in shipped code | A real risk and a common one. Search engines index it, screen readers announce it, and it reads as unfinished work. A build check that fails on known placeholder words is worth more than remembering to look. |
| Testing a translation workflow | Misleading. The expansion from English to German or from English to Chinese is not reproduced by Latin, which behaves like neither. Use real source text when the expansion ratio is the thing being tested. |
Only the opening sentence. When the Classic style is selected and the start-with option is on, the exact sentence 'Lorem ipsum dolor sit amet, consectetur adipiscing elit' is prepended to the first paragraph. Everything after it is drawn randomly from the Classic word bank, so the rest is Latin-flavoured rather than a quotation of Cicero's text. That is deliberate: it keeps every run different, which matters when you are comparing two layouts side by side.
Because it only applies to the Classic style. The condition is checked against the style as well as the toggle, so selecting Hipster, Bacon, Pirate or Corporate ignores the setting entirely rather than prefixing a translated equivalent. If you want the Latin opening, select Classic.
Because the mode generates whole sentences until the length passes the amount you asked for, then returns exactly that many characters. A hard cut is the only way to guarantee an exact length, and an exact length is the entire point of the mode. Use Words or Sentences when you want something readable, and Characters when a field has a limit you need to hit.
Placeholder text carries no rights worth worrying about, and none of these word banks is a substantial reproduction of a copyrighted work. Treat it as you would any generated filler and replace it before publishing — the practical risk is not legal but editorial, since placeholder text left in a live page is indexed by search engines and read aloud by screen readers.
No. The word banks are embedded in the page and the generation is randomised JavaScript running locally. There is no account and no saved history, and the tool keeps working with the network disconnected once the page has loaded.
Last verified: October 3, 2026
All Tools · All Tools · About · How We Test · Contact · Privacy
Maintained by Huang Bin · Last updated October 3, 2026