Text to video AI: describe one scene and its movement
Start a video clip from words when you do not have a starting image. Describe the visible scene, the subject’s main action and the camera movement, then choose a model and supported duration. For a longer story, plan several clips and check how they connect.
What this workflow helps you create
Write motion as well as appearance
A video prompt needs more than an image description. Establish the setting and subject, then state what changes during the clip. Keep the main action readable and give the camera a clear instruction such as a fixed view or a slow push-in.
Choose the model around the shot
Available models differ in supported duration, references, audio options and pricing. Choose the options that fit the shot instead of assuming every model accepts the same inputs or produces the same clip length.
Keep a longer narrative in manageable clips
A single generation follows the selected model’s clip limits. Split a sequence into an establishing shot, action and reaction when necessary. For recurring characters and dialogue across scenes, use the short drama and storyboard workflow to plan continuity.
From input to a result you can review
- 01
Plan one visible action
Choose a setting and a single main beat. Write the starting state, subject movement and intended end state. Move extra locations or complicated action sequences into additional shots.
- 02
Enter the prompt and select settings
Open the video workspace and write the scene and motion description. Select a model, aspect ratio and supported duration. Check the model’s available audio and input options, along with its pricing.
- 03
Review movement and timing
Check the subject, hands, props, camera path and ending. If the clip changes direction or rushes the action, simplify the prompt or split the beat. Download a selected result and plan the next clip around its actual ending.
Example: a text prompt with one motion goal
An empty train platform at blue hour. Warm lamps reflect on wet pavement. A loose paper ticket slides slowly across the foreground in a light breeze. Wide shot at platform level, fixed camera, subtle environmental movement, cinematic realistic style. End with the ticket resting beside a wooden bench.
This is a written prompt example. Choose a duration the selected model supports, then review whether the movement reaches a usable ending. Add dialogue or audio requests only when the chosen model supports the required output.
Choose a practical starting point
Atmospheric establishing clips
Create a quiet street, a windswept rooftop or an empty platform where environmental movement establishes the mood before a character scene.
Single-action story moments
Explore one readable beat, such as a courier turning toward a sound. Use a short test to check whether the action and camera direction support the intended story.
Visual concept and motion tests
Test a movement or atmosphere before producing a full scene. When a specific character design matters, first prepare a reference image and consider image to video.
Questions before you generate
How do I generate a video from text?
Open the video workspace, describe a scene and its movement, choose a model, aspect ratio and supported duration, then generate. Review the clip before continuing. No image is needed for a text-only prompt with a compatible model.
Can a text prompt create an entire short film?
A single request creates a clip within the selected model’s limits. A short film usually needs multiple planned shots, character references and an assembly step. Use the short drama workflow when you need scripts, storyboards and connected scenes.
Should I use text to video or image to video?
Use text to video to explore a new scene from a description. Use image to video when a reviewed character, setting or composition is an important starting point. Neither approach guarantees perfect continuity; review the generated clip before reuse.
Does text to video include sound?
Audio support depends on the selected video model and generation options. Check those options before writing a shot that relies on dialogue, music or synchronized sound. Review any generated audio together with the visuals.
Explore related workflows
New accounts receive 1,000 signup credits. Generation uses credits according to the selected model and actual usage; the bonus does not guarantee a complete video or series.