A display face is one setting, not a system
There are thirteen display families in the 150 most-used Google Fonts. Not one of them is a variable font. Eleven of the thirteen ship exactly one weight, and only two ship a real italic. Compared with the sans and monospace categories, where eight weights with matched italics is ordinary, that looks like a deficiency.
It is not. A display face is bought for one setting at one size, and the specification reflects that honestly. The useful question is not which display family has the most weights, because the answer is two, but which one survives contact with a title card. The popularity ranks agree with the reading: the best-placed display family here sits at 31, and nothing in the category cracks the top 30.
The whole category, on the fields that differ
| Family | Rank | Weights | Variable axes | Scripts | Size |
|---|---|---|---|---|---|
| Lobster Two | 31 | 2 + 2i | static | 1 | 239KB |
| Black Ops One | 43 | 1 + 0i | static | 4 | 163KB |
| Bungee | 48 | 1 + 0i | static | 3 | 116KB |
| Angkor | 71 | 1 + 0i | static | 2 | 78KB |
| Oleo Script | 78 | 2 + 0i | static | 2 | 18KB |
| Changa One | 84 | 1 + 1i | static | 1 | 24KB |
| Alfa Slab One | 86 | 1 + 0i | static | 3 | 95KB |
| Righteous | 88 | 1 + 0i | static | 2 | 42KB |
| Libre Barcode 39 | 100 | 1 + 0i | static | 1 | 14KB |
| Bangers | 102 | 1 + 0i | static | 3 | 91KB |
| Lilita One | 104 | 1 + 0i | static | 2 | 27KB |
| Titan One | 104 | 1 + 0i | static | 2 | 54KB |
| Chelsea Market | 105 | 1 + 0i | static | 2 | 80KB |
Weights counts roman styles, with matched italics after the plus. Families with one axis show its range; families with several show the first three axis names and a count of the rest, with full ranges in the profiles below. Rank is the position within the 150 families our census covers. Every name links to its own page with full licence terms.
What a single weight actually costs you
A text family with eight weights lets a designer build hierarchy inside one typeface: a heading, a subheading and a caption can all be the same face at different weights, and the page holds together because the letterforms are shared. A single-weight display face cannot do that. Every level of hierarchy beyond the one setting has to come from somewhere else.
In practice this means a display face is always a pairing decision. It carries the title and something quieter carries everything under it, which is why the sans category is ninety-five families deep and this one is thirteen. The display family is not competing with those ninety-five, it is waiting to be paired with one of them.
The two exceptions here are Lobster Two and Oleo Script, each with two weights, and Lobster Two is also one of only two families in the set with a real italic. That is a small amount of range, and it is worth noticing precisely because it is so rare in the category.
Why none of them are variable fonts
Across the other categories, variable fonts are now the norm rather than the exception. In this set of thirteen there are none at all.
The reason is economic rather than technical. A variable font requires the designer to draw the extremes of an axis and to make every position between them work, which multiplies the drawing effort for a family that will mostly be used at one weight and one size. For a text family that investment pays back across a decade of interfaces. For a face that will be set once at seventy-two points, it does not.
The practical consequence is that display faces cannot be tuned. What you see in the specimen is what you get, at every size, forever. That makes the choice of face carry more weight than it does in a text setting, where a slightly wrong weight can be nudged.
One of these is not a display face in any normal sense
Libre Barcode 39 is in this category because the metadata puts it there, and it is the smallest file in the set at 14KB. It renders text as Code 39 barcodes. It has no use in a title card and it will not read as type to a viewer, because it is not meant to.
It is left in the table rather than quietly dropped, because the category is what it is and a roundup that silently edits its own source is worth less than one that shows the outlier and says what it is. If a barcode is what is wanted, it is genuinely the tool. If not, there are twelve other families here.
Script coverage is much narrower here, and that is a real constraint
Nine of the thirteen cover extended Latin, four cover Vietnamese, one covers extended Cyrillic and one covers Khmer. Three families cover the basic Latin subset and nothing else.
This matters more for a display face than for a text face, because the display face is carrying the title, which is the one piece of text a localised version of a project absolutely has to reproduce. A title card set in a Latin-only face has no localised version; it has to be redesigned. Angkor is the only family in the set covering Khmer, and it was drawn by Danh Hong specifically for that script, which is a much stronger reason to choose it than anything about its Latin letterforms.
The thirteen, one at a time
- The largest download in the set at 239KB
- The narrowest script coverage in the set at 1 subsets
- The most roman weights in the set at 2
- The highest-ranked family in the set at number 31
- One of the few here shipping a real italic
The highest-ranked display family in the set at 31, and the closest thing here to a system rather than a setting: two weights with two matched italics, where eleven of the thirteen have one weight and no italic at all. It is also the largest download here at 239KB while covering only the basic Latin subset, which is an unusual combination and worth knowing before it goes into a project that will be localised.
- The only family here covering extended Cyrillic
- One of only 4 here covering Vietnamese
- The widest script coverage in the set at 4 subsets
A single weight, no italic, and four script subsets, which is the joint-widest coverage in this set. The coverage is the fact to carry away: for a face this specific in character, being usable across extended Latin and Vietnamese is more range than most of its neighbours offer. Credited to two designers, one of only two team credits in the category.
- One of only 4 here covering Vietnamese
One weight, no italic, three script subsets, 116KB. It was drawn by David Jonathan Ross for signage set vertically as well as horizontally, which is a genuinely unusual design brief and the reason it looks unlike the rest of this list. Nothing in the metadata captures that, and it is the clearest case in this set of a family whose point is invisible in its specification.
- The only family here covering Khmer
The only family in the set covering Khmer, drawn by Danh Hong. That single fact should decide its use entirely: if a project needs Khmer, this is the display face in the corpus that has it, and if it does not, the Latin letterforms are not the reason to come here. 78KB, one weight, no italic.
- The most roman weights in the set at 2
One of only two families here with more than one weight, at two, and at 18KB one of the two smallest downloads in the set. A connected script in a category mostly made of heavy slabs and poster faces, which makes it the odd one out visually as well as in specification. Two script subsets, so Latin and extended Latin only.
- The narrowest script coverage in the set at 1 subsets
- One of the few here shipping a real italic
One of only two families in the set shipping a real italic, alongside Lobster Two, and at 24KB one of the smallest files here. It covers the basic Latin subset and nothing else, which is the narrowest coverage in the set and rules it out of any localised work. A single roman weight, so the italic is the whole of its range.
- One of only 4 here covering Vietnamese
A single weight, no italic, three script subsets, 95KB. It sits in the middle of this set on every field measured, which is worth saying plainly: nothing in the metadata distinguishes it, and the reason to choose it is the drawing, a heavy slab serif in a category where most of the weight comes from sans forms instead.
One weight, no italic, two script subsets, 42KB, ranked 88. Like Alfa Slab One it is unremarkable in specification, which in this category is close to the norm rather than a criticism. Credited to Astigmatic, one of two names in this set credited on more than one family here.
- The smallest download in the set at 14KB
- The narrowest script coverage in the set at 1 subsets
Not a display face in the sense the rest of this list means, and the smallest file in the set at 14KB. It renders its input as Code 39 barcodes, so text set in it is machine-readable and not human-readable. It covers the basic Latin subset only. It is in the table because the category contains it and hiding it would misrepresent the corpus.
- One of only 4 here covering Vietnamese
One weight, no italic, three script subsets, 91KB, credited to Vernon Adams. A comic-lettering face, and one of the few in this set whose intended register is unmistakable from the name alone. Nothing in its specification separates it from Titan One or Lilita One; the difference is entirely in the drawing.
One weight, no italic, two script subsets, 27KB, ranked 104. One of the lighter downloads in a category where file size varies by a factor of seventeen between Libre Barcode 39 and Lobster Two. On specification alone it is interchangeable with several neighbours here, and this page will not invent a distinction that the data does not contain.
One weight, no italic, two script subsets, 54KB, credited to Rodrigo Fuenzalida, who is credited on six families across the wider corpus. Tied with Lilita One at rank 104 for the lowest position in this set. As with its neighbours, the specification is the category norm and the choice rests on the letterforms.
- The lowest-ranked family in the set at number 105
The lowest-ranked family in the set at 105, with one weight, no italic, two script subsets and an 80KB file. A hand-lettered market-sign register, which puts it closer to the handwriting category in feel than most of its neighbours here while the metadata files it as display.
Common questions
How many display fonts are in the 150 most-used Google Fonts?
Thirteen. Display is the third-largest of the five categories in that set, behind Sans Serif at 95 and Serif at 26, and ahead of Handwriting at 8 and Monospace at 7. The best-placed display family is Lobster Two at rank 31, so no display face reaches the top 30 of the corpus.
Are any display fonts variable fonts?
None of the thirteen in this set. Variable fonts are now normal in the sans, serif and monospace categories, and entirely absent here. Drawing a variable font means making every position on an axis work, which is a large investment for a face that will usually be set once at one size and one weight.
Which display fonts have more than one weight?
Two of the thirteen: Lobster Two and Oleo Script, each with two. The other eleven ship exactly one weight. That is the central practical fact about the category, because it means hierarchy beneath the title has to come from a second typeface.
Do display fonts have italics?
Almost none. Two of the thirteen ship a real italic, Lobster Two and Changa One. For the other eleven, any software asked for an italic will synthesise one by slanting the roman, which at display sizes is far more visible than it is in body text.
Are these display fonts free for commercial use?
All thirteen are under the SIL Open Font License 1.1, which permits commercial use, monetised video, broadcast, merchandise, client work and logos, with no attribution required. The licence forbids selling the font software on its own and requires a modified version to be renamed, neither of which affects using the face in a title card.
Method and sources. Category, licence and popularity rank come from our own typeface census, which reads the Google Fonts metadata API and the licence files in the google/fonts repository, and records a licence only with a source URL and a verification date. Weights, italics, variable axes, script subsets and byte size come from the Google Fonts metadata API read on the build date shown above. Popularity rank is the position within the 150 families the census covers, not Google’s global ordering. Where the metadata does not distinguish two families, this page says so rather than inventing a difference. Every family named here links to its own page on this site with its full licence terms. Built 2026-08-29.