Text Art Generators

1. ASCII banner art, made from FIGlet fonts

The oldest and most recognisable form: a word blown up to seven or eight rows tall, its letterforms assembled from smaller characters. It is what sits at the top of a project README, and what a command line tool prints when it starts.

The mechanism is lookup, not measurement. A FIGlet font is a file storing, for every letter, exactly which characters to print on each row. The generator finds your letters in that file and stitches the rows side by side, so nothing is sampled or approximated and two runs of the same word are always identical.

What it costs you is width. Every letter is several columns wide, so a ten-character phrase in a wide font can run past 100 columns and wrap badly in an 80-column terminal. Length is what breaks a banner, not font choice. The text to ASCII generator handles this family, and the ASCII text art guide goes deeper on matching a font to a destination.

2. ASCII image conversion, made by measuring brightness

Here the input is a picture rather than a word. The converter lays a grid over the image, measures the average brightness of every cell, and prints the character whose ink coverage matches that brightness. Dense glyphs like @ and # land on the dark areas, sparse ones like : and . land on the light ones, and at normal viewing distance your eye reassembles the tones into a recognisable picture.

This is why resolution matters here and not in a banner. An image conversion only carries as much detail as it has grid cells, and too few turns a face into an unreadable smear. The other lever is the character ramp, the ordered list of characters from densest to lightest, which sets the contrast curve of the whole piece. Photos go through the image to ASCII converter, footage through video to ASCII, both on the same engine as the studio, all of it local to your browser.

3. Unicode text faces and kaomoji

Text faces are small, one line, and built to live inside a sentence rather than replace it. They are assembled from characters scattered across the Unicode range, brackets, arrows, mathematical symbols and Japanese kana, chosen for their shape rather than their meaning. Most come from the Japanese kaomoji tradition, and unlike Western emoticons they are read upright, with no head tilt required.

What makes them portable is that they are a plain run of characters with no grid to preserve. A proportional font renders them fine, so they survive where a banner cannot: a display name, a bio, a post, a form field with a character limit. The one thing to watch is coverage, since a face leaning on an unusual code point becomes an empty rectangle on a device whose font lacks it. Browse them on the text faces page.

4. Glitch text, made with combining marks

Glitch text, often called zalgo, stacks combining diacritics above and below ordinary letters. Combining marks are designed to attach to the character before them, which is how accented letters work in the scripts that need them. Pile dozens onto one letter and the marks spill far outside their normal line, so the text appears to bleed into the rows around it.

Two things follow from this being real text rather than an effect. It pastes anywhere, because it is still one run of characters, which is how the look spread through usernames and bios. And it is heavy: each visible letter can carry a long tail of invisible marks, so a short glitched phrase can quietly blow past a character limit the plain version fits inside. Some platforms also strip or normalise the marks on submission, so test before you commit. The glitch text generator handles this family.