Paragraphs, sentences, words, list items, or an exact character budget, in plain text, HTML, or Markdown. Seeded so the same settings hand back the same text, and happy to build filler from your own vocabulary instead of Latin.
176 words, 1330 characters, 11 sentences generated.
Seed 20260727. The same seed always gives this exact text back.
Lorem ipsum dolor sit amet, consectetur adipiscing elit, illum est repellat, occaecat nihil dolor accusantium illo. Sunt laboris magni commodi, molestias at quo neque laboris ipsum consequat totam in. Incididunt ut anim, elit expedita consequuntur. Eveniet fugit adipisci atque, voluptates repellendus lorem laborum eos, rerum id praesentium deleniti facilis, nisi dolorum atque ea enim, sit enim eligendi nesciunt maxime esse porro distinctio voluptatibus. Minim commodi nobis doloribus elit incididunt, dolore deleniti elit. Ad explicabo itaque non fuga eaque, aute temporibus tenetur incidunt ipsum sunt voluptatem commodi, facilis aliquid suscipit laboriosam repellendus, accusamus nulla accusamus commodo nisi. Corporis dolores cillum incididunt, nesciunt ad ratione praesent aliquam, expedita maxime corporis dolore aliquip commodi vel dignissimos, maxime veniam anim. Earum optio odio adipisci consequatur pariatur laborum quibusdam occaecati ullamco aut veritatis magna sequi. Sit totam deserunt, nesciunt illum quasi est, eiusmod recusandae nisi aliquid, labore reprehenderit duis id cupidatat voluptates, eu temporibus est modi expedita neque minus provident. Commodo neque exercitation numquid, quidem quas ab blanditiis vel ducimus qui dolore voluptatum. Quas odio ullamco doloribus ullamco libero praesentium omnis.
“Neque porro quisquam est, qui dolorem ipsum quia dolor sit amet, consectetur, adipisci velit, sed quia non numquam eius modi tempora incidunt ut labore et dolore magnam aliquam quaerat voluptatem.”
The famous opening is a fragment: “dolorem” was cut after “do”, leaving “lorem”. That is why the first word is not a Latin word at all.
“Nor is there anyone who loves or pursues or desires to obtain pain of itself, because it is pain, but because occasionally circumstances occur in which toil and pain can procure him some great pleasure.”
Worth knowing before you paste placeholder into a client deck: the source is a treatise on ethics, and the sentence is about pain. It is scrambled here, so it says nothing, which is the whole point.
Once the real copy replaces the placeholder, check it against the same word, character, and reading-time budgets.
Generate at 155 characters here, then paste the real description there to see it in a search result preview.
The other half of a test fixture: identifiers to go with the placeholder text, in bulk.
Placeholder text has one job: to occupy the space real copy will take, without anybody reading it. That means the shape matters and the words do not. This generator builds sentences to a chosen length range, groups them into paragraphs at a chosen density, and draws words from the classical vocabulary the Latin was scrambled from, so the result has the rhythm of prose without the meaning. Everything runs from a seed, so the output is a function of your settings rather than of the moment you pressed the button, and every count you are likely to be asked for is recalculated as you type.
Paragraphs for body copy, sentences for a run of prose with no breaks, words for an exact count, list items for menus and cards, and characters when you are filling a field with a hard limit. Choosing the unit first is what stops you generating five paragraphs and then deleting most of them.
Short, medium, long, or mixed sentence lengths, plus paragraph density, decide whether the block reads like a product description or a legal notice. Plain text drops straight into a design tool, HTML arrives wrapped in p or li tags, and Markdown uses blank lines and dash bullets.
Every result carries the seed that produced it. Regenerate rolls a new one; leaving it alone means the same settings hand you the identical text tomorrow, which matters when a screenshot, a test snapshot, or a review comment refers to a specific line.
Meta descriptions truncate near 155 characters and a social post stops at 280, so the useful question is not roughly how long but exactly how long. Whole words stops at or under the number and rounds off with a full stop. Exact cuts precisely at the character you asked for, mid-word if that is where it lands, which is what a real field limit does to real copy.
Most generators call the random source inside the loop, so the same settings give different text every time and a layout that fitted a moment ago now overflows. Here the seed is explicit and shown next to the output. Keep it and the text is stable across reloads, browsers, and machines. Regenerate only when you want a different draft.
The classic passage is 63 distinct words, which becomes visibly repetitive past a paragraph or two. This generator draws from those plus the wider classical pool that the Latin was scrambled from, so a twelve-paragraph CMS fixture still reads like prose rather than a loop.
Paste a vocabulary and the generator builds the same sentence and paragraph shapes from it. Useful when Latin is the wrong test: a client who reads the placeholder as a mistake, a script whose glyphs are wider than Latin ones, or a search index you want filled with plausible terms rather than dolor and amet.
HTML output wraps paragraphs in p tags and list items in a real ul, and escapes any angle bracket or ampersand your own vocabulary introduces. It is ready for a template rather than something you have to clean up first.
Words, characters, characters without spaces, UTF-8 bytes, sentence count, longest sentence, and average sentence length, all recalculated as you change a setting. The byte count is the one people forget: a database column limit is usually bytes, not characters.
Nothing, and that is deliberate. It is a scrambled extract from Cicero's De Finibus Bonorum et Malorum, written in 45 BC, with words cut, reordered, and misspelled until no sentence carries a meaning. The opening is the giveaway: the source phrase is “dolorem ipsum”, and an early typesetter's cut fell after “do”, leaving the fragment “lorem”, which is not a Latin word at all. The point of nonsense is that a reader cannot read it. Real copy in a mockup pulls attention to the words, and a client who starts editing sentences has stopped looking at the layout you asked them to approve.
Google truncates on pixel width rather than character count, but around 155 characters is the practical working limit for desktop and about 120 for mobile. Generate at 155 with the whole-words option to see how much room a real sentence has, then write the actual description to fit that shape. If you are testing what happens when copy is too long, generate at 200 in exact mode and watch where your interface cuts it, because the ellipsis and the overflow behaviour are the parts that usually turn out to be wrong.
Only if it ships. Lorem ipsum on a published page is a strong signal of an unfinished site, and search engines have no way to judge a page whose text says nothing, so it competes for nothing and can drag down the perceived quality of the pages around it. The real risk is the accident: a staging page that gets indexed, a CMS template whose default value was never replaced, or a product variant nobody reviewed. Search your live site for “lorem” before a launch, and again a month later.
Both, at different stages. Placeholder is right when you are working out the layout, testing how the grid handles a long heading, or building a fixture with two hundred rows. Real copy is right before anything is approved, because copy changes design: a two-word product name and a nine-word one need different card heights, and Latin's word-length distribution is not the same as English or German. The common failure is showing a client a Latin mockup, getting approval, and discovering the real headline does not fit.
Because the classic passage only contains 63 distinct words. Any generator restricted to those will visibly loop within two or three paragraphs, and the repetition is more noticeable than in real prose because the words carry no meaning to distract from the pattern. This tool draws on the wider classical vocabulary as well, which is why longer runs still look like text. If you paste your own vocabulary, the same rule applies to it: under eight words and the output reads as an obvious loop, which the panel warns you about.
Yes. Turn on “Use my own words” and paste any vocabulary, separated by spaces, commas, or line breaks. The generator keeps its sentence and paragraph shapes and builds them from your words, so you get placeholder that looks right in a Cyrillic, Greek, or Vietnamese layout, where Latin would give you the wrong glyph widths and diacritic heights entirely. The classic opening line is disabled automatically when you supply a vocabulary, because mixing “Lorem ipsum dolor sit amet” into a Japanese sample defeats the purpose.
It is the number that produced this exact text. Placeholder is usually treated as disposable, which is fine until a colleague asks why the third card wraps onto four lines and regenerating gives you completely different text. With a seed you can hand someone the settings and the number and they see what you saw. It also makes generated fixtures stable in a test suite: the same seed produces the same rows, so a snapshot test does not fail because the filler changed.
No. The generator, the vocabulary you paste, and the text it produces all stay in this tab. There is no request to generate, no upload, and nothing stored, which is worth knowing if the vocabulary you are pasting happens to be an unreleased product's terminology.
For plain Latin text they are the same, because each of those characters takes one byte in UTF-8. They diverge as soon as the text is not plain Latin: an accented letter takes two bytes, most CJK characters take three, and an emoji usually takes four. That matters because database column limits, HTTP header limits, and many API field limits are counted in bytes, while the character count is what your interface shows the user. Both are reported here, so a vocabulary in another script tells you immediately how far apart the two numbers are.
Only in exact character mode, and on purpose. If you ask for 280 characters, you get exactly 280, including the case where the 280th character lands in the middle of a word. That is precisely what a hard field limit does to real copy, so it is the honest test of whether your interface truncates gracefully. Switch to whole words when you want text that reads properly and simply stops at or under the budget.
Explore the growing collection for calculations, documents, writing, and everyday work.