What you'll need

Step 1Set up the comp

  1. Create a new comp at 1080×1920, 30 fps, the length of your clip.
  2. Drop in your footage and scale it to fill the frame. For a two-camera podcast, cut between the speakers now, before captioning.
  3. Leave the top and bottom 250 px or so free in your head. Platforms put their buttons and captions there.

Step 2Transcribe and tidy

  1. Open Captyne (Window → Extensions → Captyne) and pick your preset on the Setup tab.
  2. Choose your transcription provider and click Transcribe.
  3. When Review Transcript opens, listen through once at 1.25×. Fix names and numbers, and delete "um" and false starts you don't want on screen. Then click Looks good — go to Preview →.
💡

Stacked blocks break at sentence ends, commas and pauses. Keep the punctuation in the transcript even if you plan to hide it on screen later.

Step 3Turn on stacked blocks

On the Preview tab, open the Stacked Blocks card next to the Interactive Preview and switch on Stacked blocks. Then use these starting values for a vertical podcast clip:

SettingStart withWhy
Words / Block min–max3 – 5Short enough to read in a glance on a phone
Line paddingSmallLines should read as one block, not separate captions
JitterAlternating, a few pxAdds a hand-set feel without looking messy
Put two short words side by sideOnStops "so I" or "to be" sitting alone on a line
Reveal ModeCumulative WordsEach block builds up as the words are spoken

Scrub through the Interactive Preview. You should already see sensible blocks. If blocks feel long, lower the maximum rather than forcing breaks.

Step 4Let AI Stacking choose

Click ✨ AI Stacking. The AI reads the whole transcript once and decides where each block breaks, how it splits into lines and which word goes big, following your Words / Block range and your preset's Emphasis tags.

When it finishes, the status line tells you how many blocks it made. Scrub the preview. If the rhythm feels off, click AI Stacking again for a different version. Running it again with nothing changed simply reuses the last result, so no tokens are spent.

Step 5Check the hero words

Read each block and ask one question: does the big word carry the meaning? Good hero words are numbers, strong verbs and surprises. Weak ones are "the", "and", "really".

Step 6Generate and polish

  1. Click Generate Captions.
  2. Select the ACP_MASTER null and open its Stack controls in Effect Controls: Line Padding, Jitter, Jitter Mode, Jitter Seed, Position and Edge Margin.
  3. Set Position so blocks sit in the middle third of the frame, clear of the platform buttons.
  4. If the jitter pattern puts two long lines on the same side, change Jitter Seed until it looks balanced.
  5. Raise Edge Margin if any long word gets close to the frame edge.
💡

Changing padding, jitter or position in the Captyne panel also updates the open comp, so use whichever you prefer.

Step 7Final checks

Make your next clip in minutes

Stacked blocks work on every plan. AI Stacking is included with Individual Pro, Studio and Enterprise.

See Plans & Pricing →

FAQ

Do I need AI Stacking to make stacked captions?

No. Stacked blocks work on every plan using automatic rules (sentence ends, commas and pauses). AI Stacking adds smarter breaks and hero-word choices.

Can I use stacked blocks with any preset?

Yes. The preset decides how normal and big words look; stacked blocks decide how they're arranged.

Why don't Words / Phrase and Words / Line change anything?

They're not used while stacked blocks are on. Words / Block takes their place.

Where's the full reference?

See Stacked caption blocks and AI Stacking in the Help Centre, and the background on the style in our blog post.