Font choice does a lot of work in short-form video. It is most of what makes captions read as "this creator" rather than "this app's default", and it is the first thing that gives away which editor someone used.

It is also more constrained on a phone than people expect, and the constraint is worth understanding before you go shopping for an app based on a font count.

Where phone fonts actually come from

Three sources, and only three.

A video frame with captions burned in, one word highlighted as it is spoken.

The system. iOS ships a substantial set of families, and any app can use them with no licensing, no download and no cost to the app's size.

Bundled with the app. An editor can ship font files inside itself. This is how apps offer the open-licence display faces people associate with viral captions. It costs app size and requires the licence to permit it.

Downloaded on demand. This is where the enormous catalogues come from. The fonts are not on your phone until you tap one, which is why they need a connection, and why the first use of one stalls.

An app claiming hundreds of fonts is almost always describing the third category. That is not dishonest, but it does mean the font is not available on a plane, and it means the catalogue is a service rather than a feature.

The listening here is Apple's Speech framework. It can run recognition on the device itself, which is why a transcript can be produced without uploading the audio.

Treat captions as access first and engagement second. The standard most teams are held to is the W3C's captions for prerecorded media.

What Zella offers, and the count

Zella on iPhone offers 46 choices, all resolved on the device. Nothing downloads, nothing stalls offline, and the catalogue costs nothing in app size beyond the handful of bundled display faces.

They are grouped so the row stays navigable rather than being one long strip:

  • Sans, the largest group at 16, including the system face and a heavy weight for titles
  • Serif, 12
  • Viral, 5, the bundled display faces short-form captions actually use
  • Handwritten, 5
  • Display, 3
  • Mono, 3
  • Script, 2

Two conveniences sit on top. Long-press a font to save it as a favourite, and favourites lead the row from then on, which matters when you use two faces for everything and there are forty-four others in the way. Favourites are stored per device rather than per project, because a preferred font is a fact about you and not about one video.

And colour accepts a hex value, typed. This is the difference between "close to our brand blue" and the brand blue, and it is the one thing eyedropper-style pickers cannot give you.

Start from a bundle, then adjust

Choosing a font, a colour, an outline, a shadow and an entrance animation one control at a time is how a caption ends up looking assembled rather than designed.

The faster route is the eight one-tap style bundles: Clean Bold, Highlight, Neon, Sunset gradient, Elegant serif, Marker, News Bar and Typewriter. Each sets the font, colour, gradient, outline, box, shadow and entrance together as one coherent look.

The useful design decision is that a bundle sets the same state the manual controls edit. It is a starting point rather than a preset you are locked inside, so picking Neon and then changing the font gives you Neon with a different font, not a fight with the app. Every raw control stays below.

The non-Latin trap

This one is genuinely worth knowing, because the failure looks like a bug rather than a font problem.

The display faces that make captions look punchy are Latin-only. They contain no Bengali, Devanagari, Arabic, CJK, Cyrillic, Greek, Thai or Hebrew glyphs at all. Force one onto text in those scripts and the system substitutes a fallback face, and depending on how the text is being drawn the result can range from "a font nobody chose" to hollow, clipped letterforms.

Zella handles this by using the system face for those scripts, which is the right answer: iOS ships proper faces for all of them and substitutes cleanly. So if you caption in Bengali and your Anton pick appears not to take, that is deliberate, and the alternative is worse.

If you are captioning in a non-Latin script, translating captions on iPhone covers the other half of that workflow.

Restraint beats variety

A practical note after all of the above: the creators whose captions look most like a brand are using one font, occasionally two. The catalogue is there so you can find yours, not so you can use forty-six.

Pick a face, save it as a favourite, set your hex colour once, and spend the attention on which caption style actually holds a viewer instead.

Frequently asked questions

How many caption fonts does Zella have on iPhone? 46 choices, grouped into seven sections, all already present on the device.

Do the fonts need to be downloaded? No. They are system families or faces bundled with the app, so they work offline and add nothing to download time.

Can I use my exact brand colour? Yes. Colour accepts a typed hex value rather than only a picker.

Can I keep my favourite fonts at the front? Yes. Long-press a font to favourite it, and favourites lead the row. They are saved per device.

Why will my chosen font not apply to Bengali or Arabic captions? Those display faces carry Latin glyphs only. The system face is used instead because it has the right glyphs and substitutes cleanly.

Can I customise a style bundle after picking one? Yes. A bundle sets the same controls you edit by hand, so anything you change afterwards simply sticks.