Describe the fictional adult subject, setting, opening composition, intended action, camera direction, pacing, lighting, and visual mood. Put the most important action early and keep it physically readable. Start with one coherent shot before adding secondary movement, environmental effects, or multiple stylistic references.
Sample

Prompt: A five-second continuous cinematic portrait of an original adult Italian woman aged 34 standing among grapevines at dusk in a fully opaque plum dress. She gently sets a bunch of grapes into a basket and looks toward the vineyard. Hair loosely tied, warm purple-green palette, restrained camera drift, fine film grain, quiet vineyard ambience.
Build the scene before the first frame exists.
Text to Video works best when the model receives a scene it can stage: who is present, what changes, how the camera moves, and what the light should communicate. Establish one opening composition and one readable action before adding atmosphere, secondary motion, or stylistic detail.

Write a brief the camera can follow
Treat the prompt as direction for one coherent shot. Keep the action readable, separate subject movement from camera movement, and then use the live catalog to choose a setup that supports the intended duration and frame.
- 01
Establish the scene
Name the fictional adult subject, setting, wardrobe, mood, and the composition you want at the opening moment.
- 02
Direct one clear movement
Describe the subject action and camera movement separately. A deliberate dolly, orbit, or tracking cue is easier to evaluate than several competing moves.
- 03
Review the live setup
Choose from the models, durations, aspect ratios, resolutions, and controls currently offered, then check the Credit estimate before submitting.
Six ways to direct a scene
These six Text to Video samples were generated with MiniMax H3 LoRA from the prompts shown. Explore portraits, travel, outdoor movement, and quiet everyday scenes. Each cover is a frame from its video; select the matching sample above to watch the motion. Your result depends on the selected model and settings.
Swipe horizontally to explore the examples.

Vineyard at Dusk
PromptA five-second continuous cinematic portrait of an original adult Italian woman aged 34 standing among grapevines at dusk in a fully opaque plum dress. She gently sets a bunch of grapes into a basket and looks toward the vineyard. Hair loosely tied, warm purple-green palette, restrained camera drift, fine film grain, quiet vineyard ambience.

A Knowing Smile
PromptA five-second continuous medium portrait of an original adult woman aged 32 in an opaque tailored jacket. She turns her head toward the viewer, smiles warmly and says "I see you looking", then gives a brief natural laugh. Gentle side light, calm camera, realistic facial movement and synchronized speech.

Forest Drive
PromptA five-second cinematic side-tracking shot of a blue rally-style coupe moving smoothly along a closed forest road in cool daylight. An original adult woman aged 35 wearing an opaque green racing jacket and a seat belt drives with both hands on the wheel. Emphasize the moving car, tree reflections and road perspective, stable vehicle geometry, realistic engine and tire ambience, one continuous shot.

Terminal Arrival
PromptA five-second smooth parallel tracking shot of an original adult brunette traveler aged 34 walking through a modern airport terminal with a rolling suitcase. She wears an opaque knee-length red dress and a tailored coat, looking ahead with a small confident smile. Bright architectural light, natural walking pace, realistic suitcase wheels, one continuous shot and quiet terminal ambience.

Morning Coffee
PromptA five-second medium shot in a cozy sunlit kitchen. An original adult woman aged 33, fully clothed in an opaque lavender sweater and tailored trousers, smiles and gently offers a white mug toward the viewer. Smooth natural hand movement, steam above the cup, warm morning light, stable framing, quiet household ambience, one continuous shot.

Meadow Run
PromptA five-second smooth tracking shot of an original adult woman aged 31 jogging at an easy pace through a green meadow with wildflowers. She wears an opaque white athletic top and relaxed running trousers. Trees and blue sky in the distance, joyful natural expression, hair moving in the breeze, realistic running motion, one continuous shot and soft outdoor ambience.
Text to Video settings for the shot
Available controls come from the selected live catalog entry. Choose duration, aspect ratio, resolution, and model around the shot you are testing instead of adding output complexity by default. Review the complete setup and Credit estimate together before submitting.
| Goal | Focus on | Why it helps |
|---|---|---|
| Validate one movement | A shorter available duration | A compact test makes it easier to judge whether the action and camera direction agree. |
| Frame for mobile viewing | 9:16 when the selected model offers it | Vertical framing keeps the subject prominent when the destination is a phone screen. |
| Show more environment | 16:9 when available | A wider frame gives tracking movement and surrounding space more room to read. |
| Control the Credit decision | Model and output configuration together | Review the live estimate after the full setup is selected, before submitting the request. |
Improve a weak video result
When a generated clip feels weak, change one prompt or setting decision at a time. Keep the model and framing stable while testing action, camera, pacing, or light so the next request shows which adjustment improved the shot.
- Too much happens at once
- Keep one main subject action and one camera movement. Move secondary atmosphere to the end of the prompt.
- The camera feels unstable
- Remove competing directions such as orbit, tracking, and zoom from the same brief.
- The subject loses emphasis
- Lead with the subject and opening composition, then simplify background activity.
- The visual mood changes
- Use one lighting scheme and one clear visual reference instead of several conflicting styles.
Text to Video questions
Practical answers about live settings, Credits, account access, and consent-first creation.
No source image is required for this workflow. The request begins with an original written concept and the settings offered by the selected model. If an exact starting composition matters more than inventing a scene from language, use an authorized source image in Image to Video instead.
Start with the shot rather than a model name. Decide whether the scene depends on duration, framing, resolution, or a particular supported control, then compare the live catalog entries that offer those choices. Review the Credit estimate after selecting the complete setup because model and output settings can affect the requirement together.
Reduce the brief to one subject action and one camera direction, then generate another controlled test. Keep the model and framing unchanged while you revise the prompt, which makes the difference easier to evaluate. Add atmosphere or secondary movement only after the central shot reads clearly.
You can inspect the generator and setup while signed out. Submission requires an authenticated account. Return to the same route after signing in.
The live Workspace shows estimated Credits for the selected model and settings before submission. Change either to see the current estimate. Catalog availability and Credit requirements can change.
Your next shot can start as one clear sentence.
For Text to Video, put the subject, action, camera, pace, and light into one focused brief, then review the live setup above.