Vizard Agent

How to make a kinetic typography video from a reference

Last updated 2026-08-16 · 5 min read

Upload the videos whose animation you want to Vizard Agent and say they define the motion language rather than the content. Vizard Agent samples frames from each reference, builds contact sheets, studies how the type actually moves, and animates your own words in that language.

What is the short version?

Kinetic typography is easy to recognise and hard to describe. Everybody can point at a video and say "like that"; almost nobody can write down what "that" is in words a tool could act on. Which is why the reference is the prompt here, and the words are almost secondary.

  1. Go to Vizard Agent and upload two or three reference videos.
  2. Say clearly that the references define the animation, not the content.
  3. Paste your script and say how long the finished film should be.

What do you need before you start?

References and a script. The references carry everything you cannot articulate — timing, easing, how words enter and leave, how much sits on screen at once — and Vizard Agent studies them frame by frame rather than describing them back to you in adjectives, which is the only way that information survives the trip.

What do you type into Vizard Agent?

Say what the references are and are not. This is the one instruction that decides whether you get your film in their style or something closer to a remake of theirs, and it is worth stating in a full sentence rather than implying it.

Prompt

Create a premium [30]-second kinetic typography film based on the attached references. The attached videos define the animation language, not the content. Script: [paste].

Variants worth knowing:

What does Vizard Agent actually do?

Here is the order Vizard Agent worked in on a real kinetic typography film. The whole opening of the run is spent taking the references apart, and it samples them more than once — the second pass is what catches the timing that a single look misses.

  1. Checks the workspace for the attached files.
  2. Loads its kinetic typography guidance rather than improvising a method.
  3. Downloads all three reference videos.
  4. Checks each reference for its technical properties.
  5. Samples frames from the references, then locates the files and samples again.
  6. Builds contact sheets from those frames so the motion can be read as a sequence.
  7. Looks at each reference in turn, reading how the type enters, holds and leaves.
  8. Animates your script in the language it has just derived.

Steps 5 to 7 are the method. A contact sheet turns motion into something that can be looked at rather than watched, which is how the easing and the rhythm get measured instead of guessed.

What does the result look like?

From the run this page is written from, probed on the delivered file: 1920x1080, H.264, 30fps, 38.73 seconds, AAC audio. Vizard Agent delivered wide rather than vertical because that is what the references were and what the prompt asked for, and it takes the shape from those two things rather than from a house default.

Ask for a vertical cut for a feed and Vizard Agent re-lays the type for the new shape rather than cropping the wide one, which matters here more than in most formats — typography that was centred for 16:9 does not survive a crop.

How long it takes was not measured separately here. Comparable work runs a median of 28 to 38 minutes end to end, and the upper quartile is several times the median. Across all projects the median cost by tier is Flash 47, Pro 55, Max 242, Ultra 263 credits.

When does this not work well?

Kinetic typography is unusually sensitive to the quality of the reference and the length of the script. Vizard Agent can match a motion language it can see clearly; it cannot rescue a film that has too many words in it.

How do you fix a result that came back wrong?

Point at the reference or at the line that is wrong. Vizard Agent keeps the contact sheets, the motion language it derived and the assembled film, so a correction adjusts the animation or the timing without sending the references back through analysis — which is the slow half of this particular job.

How does Vizard Agent compare to doing it yourself?

By hand this is After Effects, and it is genuinely skilled work: every word is keyframed, the easing is tuned by feel, and the sound has to be placed on the frame where the type lands. A thirty-second piece is days, and most of those days are spent on repetition rather than on the idea.

By hand Vizard Agent
Deriving the style Watch the reference repeatedly Contact sheets, read frame by frame
Animating each word Keyframe individually Applied from the derived language
Sound on the hits Place each effect manually Built with the motion
Changing the script Re-animate the affected lines Paste the new line

Common questions

How many references should I give it? Two or three. One works; more than one lets Vizard Agent tell what is essential from what is incidental.

Will it copy the reference's content? Not if you say the references define the animation language rather than the subject. That clause is worth including explicitly.

Can it use my brand font? Yes. Upload the font file and say to use it.

Does it write the script? It can, but this is a format where you should write it. Kinetic typography puts every word on screen, so the wording carries more weight here than in almost any other format Vizard Agent makes.

Can I get a vertical version? Yes, and Vizard Agent re-lays the type rather than cropping, which is the only way vertical typography reads properly.