Vizard Agent

How to make a deadpan text-over-clips joke video

Last updated 2026-08-30 · 7 min read

Write the list and tell Vizard Agent what should be on screen behind each line. Vizard Agent searches the stock libraries for each mundane action by name, reviews the candidates, casts a deliberately flat delivery, and cuts the text over the clips so the punchline lands on the last one.

What is the short version?

The format is a list that turns on itself: a few earnest lines, then one that undercuts them. It works because the footage is boring on purpose, which means you do not have to film anything — ordinary stock is exactly right.

  1. Go to Vizard Agent and write the list, ending on the turn.
  2. Say what should be on screen behind each line.
  3. Say the delivery should be deadpan.

What do you need before you start?

The list and a sense of the actions behind it. Vizard Agent finds all the footage, so nothing needs filming at all, but the specific mundane action behind each line is part of the joke and worth naming rather than leaving to chance.

What do you type into Vizard Agent?

Write the whole list including the ending, and name the last shot specifically. Vizard Agent builds toward it, and the final image is doing as much of the work as the final line, so leaving it unspecified wastes the punchline.

Prompt

Make a short video. Mundane footage — folding laundry, washing dishes, making the bed. Text over the clips: "[Things I thought I'd have achieved by this age:]" then [each line]. Final shot: [the one that undercuts it]. Deadpan delivery.

Variants worth knowing:

What does Vizard Agent actually do?

Here is the order Vizard Agent worked in on a real deadpan list video built entirely from stock. The searching is the interesting part: it looked for each mundane action separately by name, including the very specific final shot the joke depended on.

  1. Checks the available creative tools and the stock search options.
  2. Checks the effect presets and loads the short-form style playbook.
  3. Searches the stock libraries for the clips the list needs.
  4. Searches specifically for the final shot the punchline depends on.
  5. Searches again for that shot in a more particular form.
  6. Searches for additional mundane chore clips to fill the middle.
  7. Downloads the candidates and extracts sample frames, then reviews them.
  8. Checks the voiceover and background music options and searches the music library.
  9. Downloads the music, searches for a natural casual creator voice, then specifically for a deadpan delivery, and generates the voiceover segments.

Steps 4 and 5 are the ones that matter. The final shot is the joke, and searching for it twice — once broadly, then for the exact version needed — is the difference between a punchline that lands and one that is merely described.

What does the result look like?

A short vertical video: ordinary footage of chores, text over each clip working through the list, a flat unbothered voice reading it, and a final shot that quietly contradicts everything above it. Nothing in it was filmed by the person who made it.

The footage being unremarkable is the point. A polished shot would undercut the joke, which is why stock chores work better here than anything shot deliberately.

When does this not work well?

The format is fundamentally a joke, and jokes fail in ways that no amount of editing by Vizard Agent can rescue afterwards. These are the ones worth checking in the writing rather than discovering after the video has been built.

How do you fix a result that came back wrong?

Say which line or which shot is wrong and what it needs. Vizard Agent keeps all the downloaded candidates, the reviewed frames and the voice options, so a shot or the delivery can be swapped without rebuilding the video.

How does Vizard Agent compare to doing it yourself?

By hand this means filming yourself doing chores, which takes an afternoon to produce fifteen seconds, or searching stock and settling for whatever the first page offers. The specific final shot is the one nobody bothers to hunt for.

By hand Vizard Agent
The footage Film yourself, or settle Searched per action, by name
The punchline shot Take what you find Searched twice for the exact version
The voice Read it yourself Cast specifically for a flat delivery
Timing Adjust until it fits Cut against the narration

Common questions

Do I have to film anything? No. Vizard Agent builds the whole thing from stock footage of ordinary actions.

Does it need a voice? Not necessarily. The text can carry it, though Vizard Agent finds a flat read adds a lot.

How long should it be? Fifteen to thirty seconds. Vizard Agent keeps it tight because the joke has only one turn in it.

Can I write the list myself? Yes, and you should. Vizard Agent builds the video; the writing is the joke.

Why does the delivery matter so much? Because the humour is in the gap between the flatness of the reading and the content of the line. An enthusiastic voice closes that gap and the joke disappears with it.

Can it find very specific footage? Vizard Agent searches per action by name, which is how the punchline shot gets found at all.

What if the stock looks too polished? Vizard Agent reviews the candidates and can search again for something plainer.

Should the text be over the footage or between clips? Over it, usually. The contrast between the line and the picture is the mechanism.

Can I use my own clips instead? Yes. Vizard Agent uses your own footage exactly the same way if you have it.

Should the footage look like mine or like stock? Plain either way. Vizard Agent picks the most ordinary-looking candidates it can find, because footage that looks staged tells the viewer the whole thing is a setup before the joke has landed.

How many lines should the list have? Three or four before the turn. Vizard Agent keeps it short because the joke depends on the reader still expecting the pattern to continue when it stops.

Does music help or hurt? A quiet bed helps. Vizard Agent keeps it low and unremarkable, because anything with personality of its own competes with a delivery whose whole effect is having none.

Does Vizard Agent check the result? Yes. It reviews the clip choices and the text timing before delivery.