FIRST CH TOOLS / TEXT / 53 JAPANESE DUMMY TEXT GENERATOR

Japanese Dummy Text Generator

Filler copy for Japanese mockups and comps. Set the character count or the number of paragraphs, and drag the kanji ratio — the look of Japanese text is decided almost entirely by how much kanji is in it, so this is what makes filler behave like the real manuscript. Article mode mixes headings and bullet lists; output as plain text, HTML or Markdown; Lorem Ipsum included. Preview how it flows at any column width, font size and line height.

Set the conditions

Voice Shape Measured by

Total characters chars

10–20,000. Output stops at a sentence boundary, so the actual count lands slightly either side of the target.

Paragraph size chars

Varied by roughly ±25% per paragraph, because real manuscripts never have paragraphs of equal length.

Kanji ratio 30% Match it

Type colour in Japanese is almost entirely a function of kanji density. Working web copy sits around 25–35%, soft consumer-facing copy nearer 20%, technical documents around 40%. Untick to let the source sentences vary freely. Longer output drifts from the target — avoiding repeated sentences is given priority, so as the pool is used up the ratio pulls back toward the middle (around 30%). The figure actually achieved is in the stats below.

seed

The same seed always produces the same text. Share the number (or the whole URL) when you redraw a comp or when a team needs identical filler.

characters

blocks

avg. sentence

kanji share

Generated text Format

Count what came out with the Character Counter or the X (Twitter) Post Counter, and normalise widths with the Full-width ⇄ Half-width Converter. If you need field-level dummies — names, addresses, phone numbers, email — use the Test Data Generator instead.

Getting filler text right

Repeating あいうえお proves nothing

Filler built from one repeated character has neither the type colour nor the line breaks of real copy. What gives Japanese its texture is the alternation of kanji and kana: a run of kanji reads dark and dense, a run of kana reads pale and loose. Repeated filler has none of that variation, so the layout shifts the moment real copy arrives. That is why the kanji ratio here is a slider.

Lorem Ipsum is not a stand-in for Japanese

Latin text wraps at word boundaries and always has a space between words. Japanese wraps almost anywhere and has no spaces at all. Lay out a page with Lorem Ipsum, pour Japanese into it, and the line density, the break points and the behaviour of justified text all change. Use Lorem Ipsum when you are checking bilingual or Latin-set elements, and check Japanese boxes with Japanese.

Kanji density varies a lot by genre

Ordinary web copy sits around 25–35%, copy written for children or for a soft consumer voice around 15–22%, and technical or legal documents around 38–45%. At the same character count, 40% and 20% look visibly different in weight, so measuring the client's existing copy once and matching that number saves a round of rework. The Character Counter breaks down an existing manuscript for you.

What bites is the line count, not the character count

Cards and heroes break because the copy took one more line than expected. Characters per line is roughly column width ÷ font size: 640px at 16px is 40, a 335px phone at 15px is 22. If a card heading is designed for two lines, it breaks the moment it passes 44 characters. Use the flow preview at your real width and count the lines before you build.

Pour in the longest case and the shortest case

A design tested with one length of filler almost always fails at the edges of real data. A product name of two characters, a heading that runs to three lines, a list with a single item — all of these happen. Because you can request an exact length here, running the minimum and maximum you expect through once each removes most of the later fixing.

Check that no line starts with punctuation

Browsers apply kinsoku line-breaking rules by default and push 。、） back onto the previous line rather than start a line with them. word-break: break-all defeats that, and periods start appearing at the left margin. For Japanese, line-break: strict together with overflow-wrap: anywhere is the safe pairing. Scan the left edge of your filled layout for stray punctuation.

Justified text usually backfires in Japanese

Justification in Latin typesetting stretches the spaces between words. Japanese has no spaces to stretch, so the letter spacing itself opens up and varies from line to line. On screen, where you cannot do the manual push-in and push-out of print typesetting, left alignment is the safer default. Switch between the two in the preview and compare.

Filler that ships is the real hazard

The worst outcome for dummy text is being published. Lorem Ipsum is Latin and stands out; natural Japanese filler does not announce itself at a glance. Either mark it with a token you can grep for, or put "replace the copy" on the pre-launch checklist. The text generated here is assembled from a fixed pool of sentences and refers to no real company, person or product.

Share the seed to reproduce the same filler

Generation uses a seeded pseudo-random sequence, so the same seed and settings always produce the same text. Share the URL when you redraw a comp, when a designer and a developer need identical filler, or when you want screenshot diffs to stay meaningful. Press "Give me another" when you want a fresh one instead.

Where the sentences come from

"Business / web copy" is generic Japanese written by us and refers to no particular company or industry. "Modern literature" is drawn from works whose copyright term has expired — Natsume Sōseki (I Am a Cat, Botchan, Kusamakura, Sanshirō), Akutagawa Ryūnosuke (Rashōmon, The Spider's Thread), Dazai Osamu (Run, Melos!), Miyazawa Kenji (The Restaurant of Many Orders), Kajii Motojirō (Lemon) and Mori Ōgai (The Boat on the Takase River), all published on Aozora Bunko. The kanji usage is older than modern working copy, so corporate comps are better served by "business" or "mix".

This tool versus /testdata/

This one is for flowing prose: paragraphs, headings and bullet lists. When you need field values — names, kana readings, addresses, phone numbers, email — use the Test Data Generator, which also produces strings of exactly N characters for boundary testing. Together they cover both the rows in a list view and the body copy in a detail view.

How to Use

Set the conditionsPick a voice (business / literature / Lorem Ipsum) and a shape (paragraphs / article / one sentence per line), then enter a character count or a paragraph count. Matching the kanji ratio to the real copy is what makes the type colour ring true.
Pour it in and lookOpen the flow preview, set the column width, font size and line height you are actually using, and see how many lines you get. Characters per line is roughly width ÷ font size.
Copy it outChoose plain text, HTML (<p> / <h2> / <ul>) or Markdown and copy, or download it as a file. Share the URL when the same filler is needed again — the seed travels with it.

About This Tool

The look of Japanese text is decided almost entirely by how kanji and kana are distributed through it. Filler made of one repeated character, or of Lorem Ipsum, has a different type colour, different break points and a different line density from real copy, which is why layouts move when the manuscript lands. This tool assembles text from a pool of real sentences and weighs the kanji ratio while choosing them, so the filler can be aimed at a given density. Working web copy sits around 25–35%, soft consumer copy nearer 20%, technical documents around 40%. The requested ratio is not guaranteed — the pool is finite and not repeating sentences is given priority, so long output runs out of suitable sentences and pulls back toward the middle. The ratio actually achieved is reported in the stats, which is the number to judge by.

Sentences are never cut in half. Sentences are added until the target is reached, and the generator stops one sentence short rather than overshoot (when a sentence of exactly the right remaining length exists in the pool, it is used). Turn on "hit the character count exactly" only when you are testing a length limit — in that mode the text may end mid-sentence, and the final character is replaced with 。 to preserve the count.

Generation is seeded. The same seed and settings always produce the same text, so a comp can be redrawn or shared across a team and stay identical. There are two pools: "business / web copy" is generic Japanese written by us and names no company or industry, and "modern literature" is drawn from works whose copyright term has expired (Natsume Sōseki, Akutagawa Ryūnosuke, Dazai Osamu, Miyazawa Kenji, Kajii Motojirō and Mori Ōgai, all published on Aozora Bunko). Lorem Ipsum is available when you want to compare Latin setting.

You can see it flow before you build. Column width, font size, line height and alignment are all adjustable against live rendering, so "does the card heading stay on two lines" and "does justification open up the letter spacing" are answerable before implementation. Generation and rendering both run entirely in the browser; nothing is uploaded. Directly callable via URL parameters: /en/dummy-text/?chars=600&style=business&shape=article&format=html&kanji=30&seed=12345