To make a delayed punchline work, show the setup first, leave a short pause, then reveal the line that changes what the viewer thought they were watching. Use separate text layers in CapCut so the punchline has its own start time. The reaction shot should still make sense with the sound off.
Here is an original example: a person looks pleased while holding a blank sheet of paper. The setup reads “Me finishing my to-do list.” After a brief pause, the second line appears: “It was yesterday’s.” The expression stays in the same shot, but the text turns a small victory into a very late one.
Write the turn before choosing the clip
Start with two short lines. The first establishes an expectation; the second changes one important detail. In this example, the changed detail is which day the list belongs to. Putting both lines on screen at the beginning would reveal that detail before the viewer has accepted the setup.
A useful test is to read the setup by itself. Does it suggest a clear situation? Now read the punchline. Does it change that situation, or merely describe the same expression? If both lines mean essentially the same thing, rewrite the second line before editing.
Keep the joke original and use footage you own or have permission to use. A fictional character can carry the reaction without borrowing a recognizable person’s face or someone else’s caption.
Choose a reaction shot that leaves room for the text
Look for a continuous shot with a readable expression and a quiet background. Avoid a camera move, sign, or prop that reveals the answer before the words arrive. Leave space for the text without covering the eyes and mouth.
If you need an original reaction shot, start with CapCut’s AI video generator. A funny AI video generator can give you the visual reaction; keep the setup and punchline as editable text afterward so you can control their timing.
For this example, the shot was generated in CapCut and then edited in CapCut Web. It shows a fictional adult in a teal sweater smiling at a blank sheet before looking puzzled. The paper contains no readable answer, which lets the caption supply the turn.
An original brief for a similar shot:
Subject: A fictional adult in a plain teal sweater seated at a clean desk with a blank sheet of paper.
Action: Begin with a small proud smile. Then raise one eyebrow and give the camera a dry puzzled look.
Framing: One continuous medium close-up with a fixed camera and clear space above the face.
Keep stable: The character, clothing, desk, and lighting. Leave out text, logos, extra people, and spoken dialogue.
Review the result before adding words. Choose the moment when the expression changes, then place the punchline near that moment. If the clip already gives away the joke through a visible object or action, use another shot rather than trying to hide the answer with more text.
Give the setup and punchline separate windows
In CapCut Web, add the reaction clip to the timeline. Open Text and choose Add heading for the setup. Enter “Me finishing my to-do list,” then adjust the text size and position in Basic. Keep the entire sentence inside the frame and the face visible.
Use a second text layer for “It was yesterday’s.” You can duplicate the first layer to reuse its formatting, then replace the wording. Move the second layer to its own position on the timeline; do not leave it stacked over the setup at the opening.
The first line should be readable while the character still looks pleased. This frame from the CapCut edit shows the setup without the answer:
Trim the setup layer’s right edge so it ends before the punchline begins. Leave a short interval with neither line visible. The pause gives the viewer a moment to register the expression rather than read two sentences continuously.
For the exported example, the text windows were approximately:
These are starting points for this five-second example. A longer setup needs more reading time; a shorter reaction may need a smaller gap. Adjust the text windows to your actual words and expression instead of stretching every joke to this schedule.
Check the reveal before adding sound
Play the edit from the beginning with sound off. Ask whether the setup is clear, whether the punchline stays hidden during the pause, and whether the final line remains long enough to read. Scrub just before and just after the second text layer begins: the answer should be absent on one side and present on the other.
In the example export, the setup disappeared at about 2.07 seconds and the punchline began at about 2.57 seconds. That left roughly half a second with no text. The same shot then held the new line through the final reaction.
If the answer arrives too early, move only the punchline’s left edge later and replay the opening. If the pause feels like dead time, bring that edge slightly earlier while keeping the setup readable. If viewers miss the final line, shorten its wording or give it a longer hold.
Now play with sound on. Listen for anything that announces the answer before the text: spoken words, an effect, or a musical change. Move a revealing sound to the punchline or remove it. Keep the joke understandable without that cue; sound can emphasize the turn after the viewer can already read it.
Export and review the actual file
In CapCut Web, choose Export, then Download. Give the file a clear name and review the export settings before exporting. This example was exported as a 720p MP4 at 30 fps.
Open the downloaded file and check the opening, pause, and reveal again at normal playback speed. View it at phone size as well: the full setup must fit, the final line must be readable, and neither should cover the expression that carries the reaction.
A video meme generator can supply a reaction shot. In this example, CapCut’s AI video generator made the footage and separate text layers in CapCut Web controlled when the setup and punchline appeared. Once the joke reads clearly, explore CapCut’s AI video creation guides for another original short-video project.
FAQ
How long should I delay the punchline?
Give viewers enough time to read the setup and register the shot before showing the answer. The example uses about half a second with no text between the two lines. Adjust that gap to the words and reaction rather than treating it as a fixed rule.
Why does the joke feel flat when both lines appear together?
The second line reveals the changed detail before the setup has created an expectation. Put the lines on separate layers and move the punchline later. Replay the opening to check that only the setup is visible.
Can I make a video meme without sound?
Yes. Let the setup, expression, and final line carry the meaning. Review the edit with sound off first, then add sound only where it supports the reveal.
Should I generate the captions inside the reaction video?
For a timed text joke, add captions as separate editable layers in CapCut. That lets you change the wording and start time while keeping the same reaction shot.