You provide
A production brief
Describe the subject, setting, action, camera, light, visual treatment, and timing in the order the viewer should experience them.
No source image required
Text to video gives the model the most creative freedom. It is the right starting point when you do not have an approved frame and want to explore a subject, location, visual style, camera setup, and motion from a written brief.
You provide
Describe the subject, setting, action, camera, light, visual treatment, and timing in the order the viewer should experience them.
You direct
Choose the few details that define success. A precise subject and action matter more than a long list of decorative adjectives.
The model invents
AI decides appearance, composition, texture, depth, and all intermediate frames because no image constrains the opening.
Choose another workflow if
Start with image-to-video when a product, person, character, layout, or art direction has already been approved.
Production scenarios
Each scenario starts with a different production constraint. Use the direction note to keep the model focused on the part of the shot that matters.
Test several visual directions before committing design time to a finished frame or shoot.
Direct it: Keep the action constant and vary one dimension at a time: location, camera language, light, or visual medium.
Generate atmospheric inserts, environments, textures, and transitional footage around a larger edit.
Direct it: Specify shot size, camera movement, time of day, environmental behavior, and an ending that can cut cleanly.
Create a visually immediate opening beat for vertical social video, promos, and narrative tests.
Direct it: Put the visual change in the first sentence, use portrait framing, and avoid a setup that consumes the entire clip.
Turn a storyboard note or treatment into motion that communicates blocking, lens intent, and pacing.
Direct it: Prioritize camera and staging over surface polish so collaborators can evaluate the shot idea quickly.
Failure diagnosis
The prompt names a category but not a specific subject, action, camera relationship, or material detail.
Replace broad style words with observable choices: shot size, lens feel, movement direction, light source, texture, and timing.
Several actions and scene changes are packed into one short generation.
Treat each generation as one shot. Keep one main action, one camera move, and one visual outcome.
Text alone does not provide a fixed visual identity across frames or separate generations.
Generate or select a strong still first, then move to image-to-video when continuity matters more than exploration.
The prompt explains how motion begins but not how it resolves.
Add an ending instruction such as slows to a stop, settles on the subject, holds the final composition, or exits frame.
Include enough detail to define the subject, setting, main action, camera behavior, light, and intended finish. Remove clauses that do not change a visible production decision.
Use text to video for discovery and freedom when no source visual is fixed. Use image to video when appearance, composition, identity, or brand details must begin from an approved image.
Text can describe recurring traits, but it is not a fixed identity reference. For stronger continuity, establish the character in a still and use that image as the visual anchor for later shots.
You can also guide the video with frames — explore Framov's frame to video tools.