Video Prompting
Creating AI-generated videos requires a different approach than image prompts. This guide teaches you how to write prompts that produce compelling motion content.
Video Models in Armox
| Model | Credits | Duration | Best For |
|---|---|---|---|
| Wan 2.5 T2V Fast | 60 | 5s | Quick tests |
| Seedance 1 Lite | 300 | 5s | Short clips |
| Seedance 1 Pro Fast | 500 | 5-10s | Faster quality |
| Kling V2.5 Turbo Pro | 1,000 | 5-10s | High quality |
| Wan 2.6 T2V | 1,000 | 5-10s | Detailed motion |
| Hailuo 2.3 | 1,120 | 5-10s | Smooth motion |
| Veo 3.1 Fast | 1,500 | 5-10s | Google quality |
| Sora 2 | 2,000 | 10-20s | OpenAI quality |
| Veo 3.1 | 4,000 | 10-20s | Highest quality |
| Sora 2 Pro | 4,000 | 10-20s | Premium quality |
Text-to-Video vs Image-to-Video
Text-to-Video (T2V)
Generate video entirely from a text description.
Best for:
- Creative concepts without reference images
- Abstract or animated content
- When you want the AI to interpret freely
Image-to-Video (I2V)
Animate a starting image into video.
Best for:
- Consistent character/subject appearance
- Product animations
- When you have a specific visual in mind
- More control over the starting frame
Video Prompt Structure
Unlike image prompts, video prompts must describe motion over time:
Prompt Template
[Subject] + [Action/Motion] + [Setting] + [Camera Movement] + [Mood/Style]
Key Difference from Images
| Image Prompt | Video Prompt |
|---|---|
| "A woman standing in rain" | "A woman stands still as rain begins to fall, drops splashing on her umbrella, she slowly looks up" |
| Static description | Describes change over time |
Describing Motion
Action Verbs
Be specific about how things move:
| Vague | Specific |
|---|---|
| "Moving" | "Gliding smoothly forward" |
| "Walking" | "Striding confidently with purpose" |
| "Falling" | "Drifting down gently like a feather" |
| "Running" | "Sprinting at full speed, feet pounding" |
Motion Keywords
- Speed: slowly, quickly, gradually, suddenly, gently
- Direction: forward, backward, upward, spiraling, circling
- Quality: smoothly, jerkily, gracefully, heavily, lightly
- Intensity: subtly, dramatically, explosively, delicately
Example Motion Descriptions
The camera slowly pushes in as the dancer
gracefully extends her arms upward,
fabric flowing behind her in slow motion
Coffee pours from the pot in a smooth stream,
steam rising and curling upward,
cream swirls as it hits the dark liquid
Camera Movement
Video prompts should often include camera direction:
Camera Keywords
| Movement | Description |
|---|---|
| Push in / Dolly in | Camera moves toward subject |
| Pull back / Dolly out | Camera moves away from subject |
| Pan left/right | Camera rotates horizontally |
| Tilt up/down | Camera rotates vertically |
| Tracking shot | Camera follows moving subject |
| Crane shot | Camera moves up/down vertically |
| Orbit | Camera circles around subject |
| Static | Camera doesn't move |
| Handheld | Slight natural movement |
Camera Examples
Slow dolly in on the product sitting on the table,
maintaining focus as we get closer
Camera orbits around the model, 360-degree rotation,
keeping her centered in frame
Tracking shot following the runner through the forest,
branches passing by in foreground
Temporal Language
Describe what happens when:
Sequence Words
- "First... then... finally..."
- "As... while..."
- "Beginning with... transitioning to..."
- "Starting from... ending with..."
Example with Timing
The scene begins with an empty room,
sunlight slowly filling the space through the window.
As the light reaches the center,
a flower gradually blooms,
petals unfurling one by one
Style and Mood for Video
Cinematic Styles
| Style | Keywords |
|---|---|
| Cinematic | "Cinematic, movie-like, theatrical lighting" |
| Documentary | "Documentary style, handheld, natural" |
| Commercial | "Commercial quality, polished, advertising" |
| Music Video | "Music video aesthetic, dynamic, stylized" |
| Social Media | "Vertical format, engaging, scroll-stopping" |
Mood Keywords
- "Dramatic and intense"
- "Calm and peaceful"
- "Energetic and dynamic"
- "Mysterious and atmospheric"
- "Warm and nostalgic"
Video Prompt Templates
Product Showcase
Prompt Template
[Product] rotating slowly on [surface], [lighting type], camera orbits around showing all angles, [style] commercial quality, clean and professional
Example:
Luxury perfume bottle rotating slowly on black marble,
dramatic rim lighting with soft reflections,
camera orbits smoothly showing all angles,
high-end commercial quality,
elegant and sophisticated
Lifestyle Scene
Prompt Template
[Person/people] [action] in [setting], [time of day], [camera movement], [mood], lifestyle video, natural and authentic
Example:
Young couple walking hand in hand along the beach,
golden hour sunset, camera tracking alongside them,
romantic and peaceful mood, lifestyle video,
warm color grading, natural and authentic
Abstract/Artistic
Prompt Template
[Visual elements] [motion description], [color palette], [style], abstract and artistic, [mood/feeling]
Example:
Flowing ink drops dispersing in water,
swirling and mixing in slow motion,
deep blues and golds, abstract and artistic,
mesmerizing and meditative,
macro photography style
Social Media Ad
Prompt Template
[Product/service] [action/demonstration], [setting], dynamic camera movement, engaging and attention-grabbing, vertical format, [call to action moment]
Example:
Smartphone showcasing new camera features,
quick cuts between photo examples,
modern urban setting, dynamic handheld movement,
energetic and exciting, vertical format,
ending with product hero shot
Image-to-Video Tips
When using an image as a starting frame:
What Works Well
- β Images with clear subjects
- β Compositions with room for motion
- β Subjects in action-ready poses
- β Scenes with natural movement potential (water, fabric, hair)
What to Avoid
- β Very static, formal compositions
- β Subjects at frame edges (no room to move)
- β Complex scenes with too many elements
- β Heavily stylized images (may not animate well)
I2V Prompt Tips
Focus on describing the motion, not the image:
The woman slowly turns her head toward camera,
hair gently moving in the breeze,
soft smile forming on her face
NOT:
A beautiful woman with long brown hair
wearing a blue dress...
(The image already shows thisβdescribe the motion instead)
Common Mistakes
β Too Static
A city at night with lights
β Better
City skyline at night, cars streaming by below,
lights twinkling, clouds slowly drifting past the moon,
time-lapse feel
β Too Complex
A person walks into a room, picks up a book,
reads it, puts it down, walks to window,
opens it, looks outside, sees a bird,
the bird flies away...
β Better
A person enters a sunlit room,
pauses by the window,
gazing thoughtfully outside,
curtains gently swaying in the breeze
β Conflicting Motion
Camera zooms in while pulling back,
subject runs left while moving right
β Better
Smooth dolly in toward the subject,
who walks slowly toward camera
Model-Specific Tips
Kling
- Excellent for human motion
- Good with complex scenes
- Describe facial expressions
Veo
- Best for cinematic quality
- Handles lighting well
- Great for product videos
Seedance
- Good for dynamic motion
- Works well with dance/action
- Responsive to style keywords
Sora
- Highest quality output
- Best for complex narratives
- Understands physics well
Iteration Strategy for Video
- Start with fast models β Use Wan 2.5 Fast (60 credits) to test concepts
- Refine the prompt β Adjust motion, camera, timing
- Upgrade to quality β Once prompt is working, use premium models
- Fine-tune β Make final adjustments with high-quality model
Next Steps
- Audio Prompting β Add sound to your videos
- Video Nodes β Master video node settings
- Video Production Workflow β Complete video creation process