MiniMax H3 Video Effects: Add AE-Style Motion Graphics to Existing Footage

A finished video does not always need to be regenerated to feel dramatically different. In this House of the Dragon example, the original characters, dragon, architecture, movement, camera, subtitles, and audio remain the foundation while a new layer of red-and-white motion graphics transforms the visual rhythm.

The key is not simply asking MiniMax H3 to “add cool effects.” The prompt first defines what must stay unchanged, then assigns different effects to specific time ranges so the visual density builds toward the final dragon sequence.

Prepare the Source Video

This workflow needs:

1. An already edited source video
2. WeShop AI MiniMax H3
3. A detailed prompt divided by timeline

The input clip already contains the characters, architecture, dragon, original camera movement, subtitles, and audio. The goal is therefore not to rebuild the scene, but to treat AI effects more like a composited motion-graphics layer.

Use only footage you have the appropriate rights or permission to edit and publish.

Input

Output

Step 1: Open MiniMax H3 in WeShop AI and Lock the Source Footage

Open MiniMax H3 in WeShop AI:

https://www.weshop.ai/apps/free-minimax-h3

Use the existing video as the base footage for the edit. MiniMax H3 is presented as a multimodal workflow that can work with source video and detailed editing instructions, including requests to modify selected elements while keeping unaffected content relatively stable.

For this example, the most important part of the prompt comes before any effect description:

@Video 1 is the only base footage. Preserve the original characters, dragon, architecture, movements, camera, subtitles, and audio. Do not regenerate the scene.

This establishes several constraints at once:

• Video 1 is the only base footage.
• Keep the original characters and facial features.
• Keep the dragon anatomy.
• Keep the architecture.
• Preserve the original movement and camera content.
• Preserve subtitles and audio.
• Do not regenerate the entire scene.

When your goal is to add effects to an existing video, defining what should remain unchanged is as important as describing the effects themselves.

Step 2: Define One Visual Language for the Whole Video

After locking the source material, define the overall design direction:

Transform the footage with a luxurious, high-density, rhythm-driven AE Motion Graphics style.

The prompt then reinforces that direction at the end:

Overall style: luxurious AE Motion Graphics, high visual density, strong rhythm, red-white-black palette, fast graphic animation, particles, outlines, scanning effects, chromatic separation, motion echoes, and image slicing.

Instead of requesting one filter, this creates a consistent visual system:

red-white-black palette + fast graphic animation + particles + outlines + scanning + RGB separation + motion echoes + image slicing.

Each shot can use different effects while still feeling like part of the same edit.

Step 3: Build MiniMax H3 Video Effects Around the Timeline

The most reusable technique in this prompt is its time-based structure.

A vague instruction such as:

Add AE effects throughout the video.

does not explain how the effect intensity should change as the scene develops.

Instead, divide the footage into cuts and give each time range its own job.

CUT 1: 0:00–0:01.40 — Start Around the Character

CUT 1 (0:00–0:01.40)
Add fast white outline flashes, localized image slicing, and RGB chromatic ghosting around the character. Add short red scanning streaks across armor edges, appearing and disappearing rapidly with the motion.

The first shot introduces the style without immediately reaching maximum density.

White outline flashes, localized slicing, RGB ghosting, and short red scanning streaks concentrate the graphics around the character and armor.

This establishes the visual language while leaving room for later shots to become more intense.

CUT 2: 0:01.40–0:03.00 — Make the Graphics Follow the Architecture

CUT 2 (0:01.40–0:03.00)
Add layered geometric wireframes, perspective grids, fast scan lines, and localized duplicated image offsets around the castle. Enhance the fire area with dense particles and brief exposure flickers.

When the shot changes to the castle, the effect design changes with the subject.

Instead of repeating the character outlines, the prompt requests:

• geometric wireframes
• perspective grids
• scan lines
• duplicated image offsets
• particles
• exposure flickers

The important distinction is that the effects should respond to the existing architecture rather than replace it with a newly generated structure.

Step 4: Use Localized Effects for Close-Ups

CUT 3: 0:03.00–0:04.37 — Turn the Eye Into a Visual Focal Point

CUT 3 (0:03.00–0:04.37)
For the eye close-up, add circular scanning graphics, localized inversion flashes, RGB separation, and fast radial lines. Create one expanding graphic shockwave around the eye area without changing facial features.

For the eye close-up, the prompt becomes more localized.

Instead of applying the same effect across the full frame, it directs circular scanning graphics, RGB separation, radial lines, and a single expanding shockwave toward the eye area.

One phrase is especially important:

without changing facial features

The purpose is to decorate and emphasize the existing close-up, not redesign the character.

The same prompting pattern can be adapted when the protected subject is a face, logo, product detail, costume element, or another feature that should remain recognizable.

Step 5: Increase Effect Density With the Emotional Rhythm

CUT 4: 0:04.37–0:06.37 — Increase the Graphic Scale

CUT 4 (0:04.37–0:06.37)
Add large animated mask slices, vertical light scans, rapidly scaling geometric frames, and red-white graphic elements tracking the castle structure.

At this point, the graphics become larger.

The sequence moves from localized scanning toward:

• large animated mask slices
• vertical light scans
• rapidly scaling frames
• red-and-white architectural graphics

This raises the visual intensity without jumping directly to the climax.

CUT 5: 0:06.37–0:10.90 — Build the First Emotional Peak

CUT 5 (0:06.37–0:10.90)
Increase effect density during the shouting sequence: multi-layer motion echoes, speed lines, localized screen shake, flashing outlines, particle bursts, and expanding geometric graphics. At the emotional peak, add one strong white flash + RGB tear + circular shockwave.

The most important instruction here is not an individual effect. It is:

Increase effect density

As the character's emotional intensity rises, the motion design becomes denser as well.

The prompt combines motion echoes, speed lines, localized shake, flashing outlines, particles, and expanding geometry, then reserves a stronger combination for the peak:

one strong white flash + RGB tear + circular shockwave

Using one distinct burst at the emotional high point creates more contrast than making the entire sequence flash at maximum intensity.

Step 6: Save the Dragon Sequence for the Visual Climax

CUT 6: 0:10.90–0:15.00

CUT 6 (0:10.90–0:15.00)
Use the dragon sequence as the visual climax. Add high-speed motion trails, animated outlines, particle trails, and directional speed lines around the wings and tail. Add large geometric slicing, RGB separation, dynamic grids, and layered shockwaves. During the strongest movement, create repeated motion echoes and spatial stretching, ending with one powerful graphic burst.

The final section explicitly tells the model:

Use the dragon sequence as the visual climax.

The effect vocabulary now becomes more aggressive:

• high-speed motion trails
• animated outlines
• particle trails
• directional speed lines
• geometric slicing
• RGB separation
• dynamic grids
• layered shockwaves
• repeated motion echoes
• spatial stretching

The prompt then ends the sequence with:

one powerful graphic burst

This gives the 15-second edit a clear intensity curve: localized character effects first, architectural graphics next, a focused eye treatment, a denser shouting sequence, and finally the strongest motion treatment around the dragon.

Full MiniMax H3 AE-Style Video Effects Prompt

You can use the following structure as a starting point for your own existing footage:

@Video 1 is the only base footage. Preserve the original characters, dragon, architecture, movements, camera, subtitles, and audio. Do not regenerate the scene. Transform the footage with a luxurious, high-density, rhythm-driven AE Motion Graphics style.

CUT 1 (0:00–0:01.40)
Add fast white outline flashes, localized image slicing, and RGB chromatic ghosting around the character. Add short red scanning streaks across armor edges, appearing and disappearing rapidly with the motion.

CUT 2 (0:01.40–0:03.00)
Add layered geometric wireframes, perspective grids, fast scan lines, and localized duplicated image offsets around the castle. Enhance the fire area with dense particles and brief exposure flickers.

CUT 3 (0:03.00–0:04.37)
For the eye close-up, add circular scanning graphics, localized inversion flashes, RGB separation, and fast radial lines. Create one expanding graphic shockwave around the eye area without changing facial features.

CUT 4 (0:04.37–0:06.37)
Add large animated mask slices, vertical light scans, rapidly scaling geometric frames, and red-white graphic elements tracking the castle structure.

CUT 5 (0:06.37–0:10.90)
Increase effect density during the shouting sequence: multi-layer motion echoes, speed lines, localized screen shake, flashing outlines, particle bursts, and expanding geometric graphics. At the emotional peak, add one strong white flash + RGB tear + circular shockwave.

CUT 6 (0:10.90–0:15.00)
Use the dragon sequence as the visual climax. Add high-speed motion trails, animated outlines, particle trails, and directional speed lines around the wings and tail. Add large geometric slicing, RGB separation, dynamic grids, and layered shockwaves. During the strongest movement, create repeated motion echoes and spatial stretching, ending with one powerful graphic burst.

Overall style: luxurious AE Motion Graphics, high visual density, strong rhythm, red-white-black palette, fast graphic animation, particles, outlines, scanning effects, chromatic separation, motion echoes, and image slicing.

Do not alter character identity, facial features, original movement, dragon anatomy, architecture, or camera content. All effects must remain composited over the original footage.

You do not need to reuse every individual effect.

The transferable structure is:

lock the source footage → define the overall visual language → divide the timeline → assign effects by subject → control effect density → define a climax → repeat the preservation constraints.

Generate the Video and Review What Stayed Stable

After submitting the prompt, do not judge the result only by how many effects appear.

Compare the output against the source and check the elements you explicitly asked to preserve: character appearance, facial structure, dragon anatomy, architecture, original movement, camera content, subtitles, and audio.

In the final example, the source footage remains recognizable while red-and-white outlines, radial graphics, scanning treatments, motion accents, and stronger graphic treatment around the dragon create a distinct motion-design layer over the original sequence.

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

A video does not need to be completely regenerated to gain a different visual rhythm. Here, the transformation comes from designing effects around the original edit: beginning with restrained graphic accents, increasing density with the action, and saving the strongest visual burst for the dragon at the end.

Use the same timeline-based prompt structure with your own authorized footage, then adjust the effects, color language, and intensity for each shot in WeShop AI MiniMax H3.