MiniMax H3 resourcesBuild better video promptsOpen generator
MiniMax H3 prompt guideText to video · Image to video · Camera direction

Write prompts like compact shot briefs.

A useful video prompt is more than a mood board. Give one shot a clear subject, a readable setting, a controlled movement, a camera path, and a visual atmosphere—then remove any direction that competes with the main action.

Precise prompt formula

Subject + scene + movement + camera + aesthetic atmosphere

Write these decisions in playback order. One dominant movement and one clear camera path usually communicate more than a stack of conflicting instructions.

Subject
Who or what matters
Scene
Where the shot happens
Movement
What changes over time
Camera + mood
How the shot is seen
Basic formula
Subject + scene + motion
Precise formula
+ camera + atmosphere
Camera vocabulary
15 official commands
Prompt limit
2,000 characters
Prompt foundation

Start with three decisions. Add detail only when it earns control.

MiniMax’s basic formula is designed for open exploration. The precise formula adds professional shot direction when the framing and visual finish matter.

01

Subject

Name the main person, animal, object, or imagined form. Add appearance details only when they affect the shot.

02

Scene space

Place the subject in a concrete environment: a library, night street, studio, mountain valley, or stylized invented world.

03

Movement or change

Describe the action and how the scene evolves. Video needs a change over time, not a list of static visual tags.

04

Camera movement

Choose the framing and path: fixed, tracking, push, pull, pan, tilt, rise, fall, zoom, or a restrained combination.

05

Aesthetic atmosphere

Set lighting, palette, texture, realism, weather, lens feeling, and emotional tone without overwhelming the action.

06

Final beat

State where the shot should settle: a close expression, a wide reveal, a completed turn, or a still product frame.

The two formulas

Basic: subject + scene + movement/change. Precise: subject + scene + movement/change + camera movement + aesthetic atmosphere.

Reusable templates

Choose the input mode before writing the first line.

Text-to-video must establish the whole visual world. Image-to-video can spend more of the prompt on motion because the first frame already defines subject and style.

Text to video

Scene-building template

Use when the model must invent the subject, environment, and opening composition from text.

T2V template
[Subject and defining appearance] in [specific scene and time]. [Primary action or change]. [Camera command or natural-language camera path]. [Lighting, color, visual style, and atmosphere]. The shot ends on [final composition or emotional beat].
Image to video

Motion-first template

Use when the uploaded first frame already establishes identity, composition, and the initial aesthetic.

I2V template
The [main subject visible in the image] [performs one clear action]. [Environment or secondary elements change]. [Camera path] while preserving the subject’s defining appearance and the original visual style. End with [final pose or framing].
First and last frame

Transition template

Describe how the first controlled image evolves into the last rather than re-describing both frames in isolation.

FL2V template
Beginning from the first frame, [subject action and environmental transition]. [Camera movement and timing]. Preserve [identity, object, or style constraints] as the composition evolves naturally into the supplied final frame.
Subject reference

Performance template

Keep appearance references separate from the action, camera, and environment that should change.

S2V template
The referenced character appears in [scene] wearing [relevant wardrobe]. They [single readable action] with [controlled expression]. [Camera path] reveals [important detail]. Keep facial identity consistent and finish on [final framing].
Camera vocabulary

All 15 documented English camera commands.

Use the bracketed form for supported models when camera accuracy matters. Natural-language direction also works, but explicit commands provide a clearer control signal.

Horizontal travel

Truck

[Truck left][Truck right]

Move the camera laterally while preserving its viewing direction.

Horizontal rotation

Pan

[Pan left][Pan right]

Rotate the camera left or right from a relatively fixed position.

Forward and backward

Push and pull

[Push in][Pull out]

Physically move toward or away from the subject to change spatial emphasis.

Vertical travel

Pedestal

[Pedestal up][Pedestal down]

Raise or lower the camera without tilting its viewing direction.

Vertical rotation

Tilt

[Tilt up][Tilt down]

Rotate the view upward or downward from a relatively fixed position.

Lens framing

Zoom

[Zoom in][Zoom out]

Change apparent focal length rather than physically traveling through the scene.

Energy

Shake

[Shake]

Introduce deliberate camera instability; use sparingly so subject motion remains readable.

Subject-led path

Follow

[Tracking shot]

Move with the subject to preserve their position while the environment changes.

No camera travel

Static

[Static shot]

Hold the camera steady and let performance or environmental motion carry the clip.

Control rules

Sequence the camera instead of stacking contradictions.

Commands are most useful when they clarify the shot’s order and priority.

Combined movement

Keep one bracket group below three commands

Commands inside one bracket group act together, for example [Pan left,Pedestal up]. More than three is not recommended.

Sequential movement

Place commands where they happen

Write the action in order: the subject enters [Tracking shot], pauses, then the camera reveals the face [Push in].

Short shot

Keep complex choreography within 5–6 seconds

MiniMax’s prompt guide recommends a compact movement window to reduce failures from overly complex camera direction.

Prompt optimizer

Choose interpretation or exact wording

The API defaults prompt_optimizer to true. Disable it when the precise command sequence should be preserved.

Example prompts

Four shot briefs you can reshape.

These examples are newly written from the documented formula. Replace the subject and setting while preserving the hierarchy of direction.

Text to video · tracking

Rain-lit cyclist

One subject action, one camera path, and a final emotional beat.

6-second cinematic shot
A courier in a reflective jacket cycles through a rain-lit city intersection at night. Water sprays from the tires as neon reflections stretch across the road [Tracking shot]. The camera moves beside the bicycle, then gently closes on the courier’s focused expression [Push in]. Natural wet-surface reflections, cool blue light with restrained magenta accents, cinematic realism.
Image to video · product

Watch reveal

Let the first frame own the product design while the prompt directs light and camera motion.

Controlled image animation
The watch remains centered as a narrow band of warm light travels across the brushed metal case. Fine mist moves slowly behind the product [Truck right]. Preserve the watch geometry and dark marble surface. End on a clean three-quarter product angle with the dial fully readable.
Text to video · static

Quiet performance

Use a fixed camera when expression is more important than spectacle.

Character close shot
A stage actor sits alone at a dressing-room mirror after a performance. Their shoulders relax and a restrained smile appears as distant applause fades [Static shot]. Warm practical bulbs frame the reflection, soft film grain, natural skin texture, intimate and contemplative atmosphere.
Sequential movement

Forest reveal

Put each camera command beside the moment it should control.

Two-part camera path
A red fox steps through a misty cedar forest at dawn [Tracking shot]. It stops beside a shallow stream and looks upward as sunbeams break through the canopy. The camera slowly rises to reveal the scale of the forest [Pedestal up]. Natural color, cool mist, soft golden light, calm documentary realism.
Common failure patterns

Remove ambiguity before adding adjectives.

Most weak prompts do not need more words. They need a clearer subject, action, or camera priority.

Keyword cloud

Static style tags with no change over time

Replace “cinematic, beautiful, 4K, dramatic” with a subject performing one visible action inside a defined scene.

Competing directions

Several camera moves at the same priority

Choose one dominant movement. If a second is necessary, place it later in the prompt as a clear sequence.

Overloaded story

Too many beats for one short clip

Give the shot one job. Build a longer sequence from several generated clips rather than compressing a full scene into six seconds.

Image repetition

Re-describing everything already visible

In image-to-video, spend prompt space on movement, change, camera behavior, and the features that must stay consistent.

Unclear ending

The shot has no final composition

Name the last beat so the action resolves: a close expression, a wide reveal, a still pose, or a completed transition.

Optimizer mismatch

Expecting exact commands after automatic rewriting

When precise bracket order matters, test prompt_optimizer false and compare the result against the default optimized request.

Before generation

A six-question prompt check.

Read the prompt once as a director and once as a timeline.

Subject

Is one subject clearly dominant?

A viewer should know what to watch before the camera starts moving.

Scene

Is the environment concrete?

Name place, time, weather, or production context when it changes the visual result.

Movement

Does something visibly change?

Describe an action, transformation, expression, or environmental event across time.

Camera

Is there one primary camera path?

Use a documented command or a concise natural-language path, not both repeatedly.

Atmosphere

Do light and style support the action?

Keep the look coherent and avoid mixing several unrelated aesthetics.

Ending

Can you picture the final frame?

A clear last beat helps the model resolve motion instead of drifting into a random continuation.

Prompt questions

Direct the important part first.

Should I write a long prompt?

Length is not the goal. Use enough detail to identify the subject, setting, movement, camera path, and atmosphere, then remove instructions that compete with the primary action.

Can I combine camera commands?

Yes. Commands inside one bracket group act together, such as [Pan left,Pedestal up]. MiniMax recommends no more than three commands in one combined group.

How do I sequence two camera movements?

Place each command beside the relevant action in playback order—for example, follow the subject [Tracking shot], then reveal an expression [Push in].

Should prompt_optimizer be enabled?

It defaults to true and can improve broad creative prompts. Set it to false when you want the API to preserve exact wording or a deliberate camera-command sequence.

How is image-to-video prompting different?

The first frame already establishes subject, composition, and style. Focus the prompt on motion, camera behavior, environmental change, and the features that must remain consistent.

From brief to motion

Write one clear shot. Then test the direction.

Start with the basic formula, add one useful camera decision, and compare the result before increasing complexity.

MiniMax H3

A focused, independent video-generation experience built around clear creative direction.

MiniMax H3 is an independent product experience and is not affiliated with or endorsed by MiniMax. This site does not connect to an official MiniMax H3 API, and checkout is not available.

© 2026 MiniMax H3
MiniMax H3 Prompt Guide | Hailuo Video Prompt Templates