Approachably
← All posts
Craft

How a 15-second cartoon Short gets made

Oct 21, 2026·4 min read
#how to make a cartoon short#youtube shorts#animation#duma meow#ai voices

A 15-second cartoon Short gets made in a handful of stages: pick a format the audience already understands, write three or four beats, record and trim the voice lines, animate to the audio, run safety checks and give it a clear title. That's how I make Duma Meow, a Vietnamese 2D cartoon comedy about four cats, and the same steps work for almost any short cartoon.

Fifteen seconds sounds tiny, but every frame has to earn its place. Here's how each stage works.

Step 1: Choose a format people already know

The fastest way to land a joke in 15 seconds is to borrow a shape the viewer already recognizes. Duma Meow skits use formats like:

  • the side-eye reaction
  • this is fine, calm in a disaster
  • the expanding brain
  • the starter pack
  • balloon can't drop
  • zoomies
  • dance trends such as a tiptoe kick-strut or a flash-dodge dance

When the format is familiar, you don't spend seconds explaining the setup. The audience fills it in, and you spend your time on the twist.

Step 2: Write beats, not scenes

A Short has room for roughly three beats: setup, escalation, payoff. Sometimes a fourth, a tiny button after the punchline, if it fits.

Character does a lot of the work. Duma Meow has four cats with clear personalities: Maelis is deadpan, Corin is flustered, Seraph is the wildcard and Luna is always prepared. If you know who reacts how, a beat can be as small as Maelis saying Haizz (a sigh) while chaos happens around her.

Keep dialogue short. Lines like Ủa? ("Huh?"), Chờ chút ("Wait a sec") or Mười điểm ("Ten points") read instantly and leave room for the visuals. For more on that, see writing jokes for cats who barely talk.

Step 3: Make the voices

Duma Meow is in Vietnamese, and the voices are generated with ElevenLabs v3. Each cat has a provisional voice. Before settling on them, I ran a casting test where every cat said Ham chơi! in every voice, which you can watch in the ham chơi voice test. The full story is in how I cast voices for four cartoon cats.

Once lines are generated, each one is trimmed, never time-stretched. Stretching audio to fit a gap makes speech sound warped, and in a tonal language like Vietnamese it can blur the pitch that carries meaning. If a line is too long, I rewrite the line or change the timing of the animation instead.

Step 4: Animate to the audio

With the audio locked, animation follows the sound. Two things matter most in a Short:

  • Timing. The visual punchline should hit on the beat of the audio, not near it.
  • Mouths. In Duma Meow, mouth flaps are driven by loudness: the louder the line at that moment, the wider the mouth opens. It's a simple approach that suits a 2D cartoon style and keeps lip movement tied to the real voice.

Everything is built for vertical 1080×1920 video, since that's the native shape for Shorts. Framing for vertical means keeping characters centered and big; there's no room for wide establishing shots.

Step 5: Run safety checks

Every Duma Meow clip passes a flashing-safety check before it goes out, and the target is 0 flashes per second. Rapid flashes can be a problem for viewers with photosensitive conditions, and cartoon effects like explosions or camera flashes are easy places for them to sneak in. Checking each clip, every time, is cheaper than fixing a problem after upload.

The other standing rule: the cats are cartoon characters, and slapstick stays cartoonish. In skits like the slap-off challenge, no cats are harmed.

Step 6: Title it in two languages

A Short lives or dies by whether the right person finds it. Duma Meow titles are written in English and Vietnamese, so a Vietnamese speaker and an English speaker searching for the same trend can both land on it.

Good titles name the format or the joke plainly. Clever titles that only make sense after watching tend to get skipped.

What makes 15 seconds work

  • One idea per Short.
  • A format the viewer already knows.
  • Short, clear lines that are trimmed, not stretched.
  • Movement that lands on the audio.
  • A safety check every single time.

You can see the results on the Voiceover with jimwest405 channel. If you're making your own, start by writing just the punchline, then work backward to the shortest possible setup.

More posts