Stacked Text Generator

Stacked text needs two inputs: the normal line that stays at full size, and the small line that floats above it.

1. Type your main line into the Bottom text box. It can contain any characters, including digits, spaces and emoji.
2. Type the small words into the Top text (small letters stacked above) field.
3. Read the result in the Stacked text box. If any top letters could not be used, a note under the box says how many were skipped and lists the letters that are supported.
4. Press Copy or Download .txt. Use Clear to empty the bottom box and start again.

With the bottom text welcome and the top text hi, the result is wͪeͥlcome: the small h sits on the w and the small i sits on the e.

The effect relies on combining characters. A combining mark has no width of its own and attaches to the character before it, the way an accent sits on a vowel. Unicode includes a set of combining marks drawn as miniature Latin letters. The generator places the first top letter after the first bottom character, the second after the second, and so on.

• Position matters. The top text is not centred over the word. Each small letter lands on the bottom character in the same position.
• Spaces are gaps. A space in the top text leaves that bottom character bare, which is how you shift the small word to the right.
• Unsupported letters are skipped, not shifted. The other letters keep their positions. With bottom text hello world and top text joy, the j and y are skipped and the o lands on the e, giving heͦllo world and a note that two letters were skipped.
• Extra letters are dropped. If the top text is longer than the bottom text, the surplus is discarded without a warning.
• Case is ignored. Capitals in the top text are treated as lowercase.

To plan a pairing, write the top word under the bottom line and count along. With the bottom text WELCOME HOME and the top text hello, the five small letters land on W, E, L, C and O, which gives WͪEͤLᷝCᷝOͦME HOME. The small word sits over the first half of the first word, not over the whole phrase. If you want it over HOME instead, begin the top text with eight spaces so that the first small letter reaches the H, which is the ninth character.

Twenty-three letters have a combining form: a, b, c, d, e, f, g, h, i, k, l, m, n, o, p, r, s, t, u, v, w, x and z. There is none for j, q or y, and none for digits or punctuation, so those are skipped. Thirteen of the letters (a, c, d, e, h, i, m, o, r, t, u, v and x) come from the long-established Combining Diacritical Marks block and are widely supported. The other ten (b, f, g, k, l, n, p, s, w and z) come from a later supplement block, and plenty of fonts do not draw them.

• Use one line. Positions count line breaks as characters, so a multi-line bottom text shifts the pairing.
• Keep the top short. Three to eight letters read clearly. Longer phrases blur into a smudge.
• Match the lengths. A bottom line of similar length to the top line looks tidiest.
• Test the destination. Some apps clip the top of a line or compress the marks, so paste a trial first.
• Do not rely on it. Screen readers announce combining marks oddly, so keep important information in ordinary text.

For a smaller single line, see the small text generator or the superscript generator. The Zalgo glitch text generator uses combining marks in a wilder way, and the Unicode text converter shows the code point of any mark you paste.