AI Text Animation with MiniMax H3: Create a Black-and-White Kinetic Typography Video

A word flashes onto a black screen, expands until it nearly fills the frame, then breaks apart. Seconds later, rows of type collide, multiply, invert, and rebuild themselves into a new composition.

There are no people, photos, products, or background scenes. The typography itself is the entire visual.

That distinction is what separates kinetic typography from a conventional video with captions. Instead of asking AI to “add animated text,” this workflow defines the exact words, the motion assigned to them, and everything the model should keep out of the frame.

This tutorial uses WeShop MiniMax H3 to create a black-and-white AI typography motion video with a small amount of dark red accent.

1. Make Typography the Entire Visual

The first challenge is preventing the result from becoming a normal scene with captions layered on top.

Start by defining a strict visual boundary:

The entire video is a dynamic typography piece on a pure black background.

Only text itself is allowed to appear in the frame.

Do not show people, landscapes, photos, illustrations, icons, products, logos, 3D objects, UI elements, timers, borders, grids, or decorative graphics.

The most important instruction is:

Only text itself is allowed to appear in the frame.

This tells the model that typography is not an information layer placed over another visual. Typography is the visual subject.

The original concept establishes typography as the only visual subject.

2. Define the Exact Words Before Defining the Motion

“Create kinetic typography” tells the model how the video should feel, but not what should actually appear on screen.

For this experiment, the text is planned in advance.

Opening

TEXT
MOTION
TYPE
EXPERIMENT

These words immediately establish the visual theme.

Speed and Transformation

FAST ENTRY
INSTANT SCALE
RAPID ROTATION
COMPRESS & REBUILD

The language itself describes motion, which makes it easier to pair each phrase with a visible action.

Layout

MULTI-LINE LAYOUT
SIZE CONTRAST
OFFSET STRUCTURE
RHYTHM SHIFT

This group shifts the composition from isolated words into denser graphic layouts.

Strongest Motion Sequence

COLLISION
DECONSTRUCT
MULTIPLY
INVERT
REBUILD

Each word receives its own behavior:

• COLLISION — text rushes toward the center from different directions.
• DECONSTRUCT — complete words break into letter fragments.
• MULTIPLY — one word rapidly becomes many copies.
• INVERT — the composition flips or mirrors.
• REBUILD — scattered elements return to an organized composition.

Visual Character

BLACK
WHITE
SHARP
CALM
RESTRAINED

These words slow the rhythm temporarily while reinforcing the visual language of the piece.

Controlled Color Distortion

COLOR SHIFT
TEXT OFFSET
MOMENTARY GHOST
LAYER OVERLAP

This section introduces a very small amount of dark red as an accent without allowing it to replace the black-and-white visual system.

Closing Sequence

TEXT IS MOVING
LAYOUT IS CHANGING
STRUCTURE IS REBUILDING

The final title becomes:

AI TYPOGRAPHY MOTION

Planning the actual copy gives the animation a progression instead of leaving every text choice to the model.

3. Connect Each Word to a Visible Action

Abstract style words such as “dynamic,” “premium,” or “experimental” can describe the mood, but they do not explain what should move.

For AI typography motion, verbs are more useful.

The prompt uses actions including:

• scaling;
• shrinking;
• rotation;
• flipping;
• compression;
• stretching;
• collision;
• deconstruction;
• multiplication;
• offset;
• overlap;
• mirroring;
• rearrangement;
• rebuilding.

For the strongest sequence, the relationship is explicit:

“COLLISION” — multiple text blocks rush toward the center from different directions and collide.

“DECONSTRUCT” — complete words break apart into fragments or partial letter structures.

“MULTIPLY” — one word instantly duplicates into many copies that spread across the frame.

“INVERT” — an entire group of words flips vertically, horizontally, or mirrors briefly.

“REBUILD” — scattered text rapidly regroups into a clean composition.

Instead of simply requesting “cool animated typography,” this gives the model an object, an action, and an intended visual result.

4. Protect the Minimal Black-and-White Style

A typography prompt can easily drift toward neon graphics, HUD interfaces, particles, or other decorative effects.

This experiment needs the opposite: a restrained 2D graphic-design language.

The negative constraints are therefore explicit:

Do not use cyberpunk style.
Do not use neon lights.
Do not use glowing text.
Do not use particles.
Do not use smoke.
Do not use HUD graphics.
Do not use real-world objects.
Do not use people.
Do not use illustrations.
Do not use photos.
Do not use logos.
Do not use Chinese text.
Do not use numbers.
Do not use random gibberish.

The prompt also keeps the composition visually flat:

Keep a flat 2D graphic design feeling.
No cinematic camera movement.
No scene perspective.
No realistic depth of field.

The goal is a motion-design composition made from typography, not a cinematic scene that happens to contain text.

5. Copy the Complete AI Typography Motion Prompt

Use the full prompt below:

AI typography motion, 16:9 landscape, high-density kinetic typography experimental short video.

The entire video is a dynamic typography piece on a pure black background. Only text itself is allowed to appear in the frame. Do not show people, landscapes, photos, illustrations, icons, products, logos, 3D objects, UI elements, timers, borders, grids, or decorative graphics.

Use only the following English text in the video. Do not generate unrelated words:

Group 1:
TEXT
MOTION
TYPE
EXPERIMENT

Group 2:
FAST ENTRY
INSTANT SCALE
RAPID ROTATION
COMPRESS & REBUILD

Group 3:
MULTI-LINE LAYOUT
SIZE CONTRAST
OFFSET STRUCTURE
RHYTHM SHIFT

Group 4:
COLLISION
DECONSTRUCT
MULTIPLY
INVERT
REBUILD

Group 5:
BLACK
WHITE
SHARP
CALM
RESTRAINED

Group 6:
COLOR SHIFT
TEXT OFFSET
MOMENTARY GHOST
LAYER OVERLAP

Group 7:
TEXT IS MOVING
LAYOUT IS CHANGING
STRUCTURE IS REBUILDING

Ending text:
AI TYPOGRAPHY MOTION

All visible text should be in clear English uppercase letters. Avoid gibberish, random symbols, distorted unreadable letters, Chinese characters, Japanese characters, or any unrelated language.

The main visual subject is white typography on a pure black background, creating strong contrast. Only a very small amount of dark red text is allowed for brief offset, emphasis, or visual impact. Dark red should never dominate the frame.

The text is not a subtitle layer and not static layout. The text itself is the only protagonist of the video.

Start from a pure black screen.

The word “TEXT” suddenly appears in the center in white, then scales up rapidly until it fills most of the frame.

Then “MOTION / TYPE / EXPERIMENT” appear in quick succession and rearrange horizontally and vertically in the center.

Move into a denser typographic section with:

“FAST ENTRY”
“INSTANT SCALE”
“RAPID ROTATION”
“COMPRESS & REBUILD”

These phrases switch rapidly. The text changes size dramatically and moves quickly across the frame from left, right, top, and bottom.

Then show:

“MULTI-LINE LAYOUT”

This phrase is duplicated into multiple rows that fill the screen.

Then switch to:

“SIZE CONTRAST”

Use one oversized word together with multiple very small text lines to create an extreme scale difference.

Then show:

“OFFSET STRUCTURE”
“RHYTHM SHIFT”

The text becomes horizontally and vertically misaligned. Different text blocks push, collide, and rearrange.

Then enter the strongest action section with:

“COLLISION”
“DECONSTRUCT”
“MULTIPLY”
“INVERT”
“REBUILD”

Each word should match its own motion behavior.

“COLLISION” — multiple text blocks rush toward the center from different directions and collide.

“DECONSTRUCT” — complete words break apart into fragments or partial letter structures.

“MULTIPLY” — one word instantly duplicates into many copies that spread across the frame.

“INVERT” — an entire group of words flips vertically, horizontally, or mirrors briefly.

“REBUILD” — scattered text rapidly regroups into a clean composition.

Then the rhythm becomes more restrained.

Show these words one by one in the center:

“BLACK”
“WHITE”
“SHARP”
“CALM”
“RESTRAINED”

Each word should appear in large, minimal typography with fast but controlled switching.

Then introduce a very small amount of dark red text with:

“COLOR SHIFT”

The main white text remains stable while a dark red duplicate briefly shifts sideways.

Then show:

“TEXT OFFSET”
“MOMENTARY GHOST”
“LAYER OVERLAP”

Red is only a brief accent and never becomes the main color.

Finally, show the phrases:

“TEXT IS MOVING”
“LAYOUT IS CHANGING”
“STRUCTURE IS REBUILDING”

These three lines appear in multi-line form, rapidly changing size, spacing, and position, then gradually converging toward the center.

All text disappears quickly.

The final words:

“AI TYPOGRAPHY MOTION”

They rapidly rebuild from scattered fragments, stabilize in the center on the pure black background, stay briefly, then cut to black.

Overall style:
high-end, restrained, sharp, rational, inorganic, experimental.

The result should feel like a designer-made kinetic typography / typography motion showreel.

Use pure black background, pure white typography, and only a minimal amount of dark red accent.

Text motion should include:
scaling, shrinking, rotation, flipping, compression, stretching, collision, deconstruction, multiplication, offset, overlap, inversion, mirroring, rearrangement, and rapid rebuilding.

The pacing should be fast but still controlled.

Every text group should remain clearly readable for at least a brief moment.

Keep a flat 2D graphic design feeling. No cinematic camera movement, no scene perspective, no realistic depth of field.

Do not use cyberpunk style.
Do not use neon lights.
Do not use glowing text.
Do not use particles.
Do not use smoke.
Do not use HUD graphics.
Do not use real-world objects.
Do not use people.
Do not use illustrations.
Do not use photos.
Do not use logos.
Do not use Chinese text.
Do not use numbers.
Do not use random gibberish.

This is not a subtitle video. It is a high-density experimental motion piece built entirely from typography itself.

6. Generate the Video with WeShop MiniMax H3

Open WeShop MiniMax H3.

MiniMax H3 supports prompt-led video workflows on WeShop, so the detailed typography direction can be supplied directly as the creative brief.

Paste the complete prompt into the generation interface and start the generation.

There is no need to introduce a person, product, photograph, or environment for this particular workflow. The typography itself provides the content and motion.

When the result is ready, check three things:

1. Is typography still the dominant or only visual subject?
2. Have unwanted people, images, logos, decorative graphics, or unrelated text appeared?
3. Does the typography actually scale, collide, deconstruct, invert, and rebuild instead of behaving like simple captions?

If the output starts looking like an ordinary subtitle video, reinforce these two lines in the next prompt:

Only text itself is allowed to appear in the frame.

This is not a subtitle video. It is a high-density experimental motion piece built entirely from typography itself.

7. Review the Final AI Text Animation

The final result should read as a sequence of changing graphic compositions rather than a conventional video with text placed on top.

The black background keeps the visual field empty while white typography carries the rhythm. A small amount of dark red can interrupt the composition briefly, while scaling, repetition, collision, inversion, and reconstruction provide the movement.

You can try it directly using the tool on the right.

Turn Text Into the Visual

The interesting part of this approach is how little the frame actually needs. Remove the scene, the subject, and the decorative layer, and the words themselves can carry scale, rhythm, impact, and transformation.

Once the copy, motion verbs, and visual restrictions are defined together, text stops behaving like a caption and starts behaving like motion design.

Create your own AI typography motion with WeShop MiniMax H3