Write the shot before you write a long prompt
Use the text-to-video page when the source material is still an idea. Build the scene as a short shot brief, choose an available model, then compare motion, pacing, and framing before spending more credits on final variations.
A cleaner way to prompt
Instead of stacking every adjective into one paragraph, treat each generation like a small production note.
Frame the intent
Name the audience, channel, and job of the clip: product teaser, cinematic opener, social hook, or storyboard beat.
Describe the shot
Write the subject, action, environment, lens movement, and timing in direct language that a director could follow.
Judge the first pass
Check whether the motion reads correctly before you chase texture, lighting, or brand-specific styling.
Change one variable
Keep the winning structure and test one change at a time: camera move, mood, product angle, or opening beat.
What belongs in this page
Net-new video concepts
When there is no source image yet, text-to-video is the fastest route to a rough scene and motion direction.
Identity-critical subjects
If a person, product, or exact visual frame must stay recognizable, start from image-to-video instead.
Short, testable briefs
A compact brief with clear camera logic usually beats a dense paragraph of unrelated style requests.
Practical text-to-video jobs
Ad concept rough cuts
Create opening hooks and benefit shots before choosing which direction deserves a finished asset.
Storyboard motion tests
Turn still storyboard language into moving beats that a team can review quickly.
Social clip exploration
Draft several pacing and camera options for short-form channels without starting from a reference image.
Text-to-video questions
How is this different from the main AI video generator?
The main generator is a broad workspace. This page is intentionally narrow: it is for projects that start with a written scene and no required visual reference.
What should my first prompt include?
Include the subject, action, setting, camera movement, mood, and output goal. Leave brand refinements for later passes.
When should I switch to image-to-video?
Switch when a source photo, product, character, or existing frame should guide the generated result.
Start with a shot brief, not a wall of text
Write one clean scene, generate a first pass, then keep the strongest motion logic for the next variation.