Input Controls

3.1 Uploading an Initial Image

You can upload an initial image to guide video creation.Purpose:

  • Animate a static image

  • Add motion effects

  • Provide visual context to the AI

The AI uses this image along with your prompt to create a dynamic sequence.


3.2 Uploading a Final Image (Model-Dependent)

Some models (such as Luma Labs, Kling 2, Quan V1) support uploading a final image.This allows:

  • Smooth animation from the first image to the last

  • Seamless transformation sequences

  • Cinematic transitions

If supported by the selected model, both initial and final images can be used to control narrative progression.


3.3 Writing the Prompt

You can enter a descriptive text prompt (maximum 3,000 characters).The prompt defines:

  • Scene

  • Action

  • Environment

  • Camera behavior

  • Lighting

  • Mood

  • Style

Well-structured prompts produce higher-quality outputs.


3.4 Magic Prompt (AI Enhancement)

The Magic Prompt feature automatically enhances your prompt.It:

  • Refines language

  • Adds cinematic detail

  • Optimizes prompt structure for the selected model

  • Generates improved scene descriptions

This is particularly useful for users seeking professional-level outputs without manual refinement.


3.5 Cinematic Keyword Controls

On the left panel, you can apply structured visual modifiers to guide generation.These are organized into categories:

Motion

Examples:

  • Static

  • Move left

  • Move right

These define camera or subject movement.


Lens

Examples:

  • Zoom lens

  • Infrared lens

  • Fisheye lens

These influence perspective and visual style.


Shot Size

Options such as:

  • Close-up shot

  • Medium shot

  • Long shot

These define framing and subject distance.


Lighting

Examples:

  • Studio lighting

  • Soft lighting

  • Cinematic lighting

Lighting impacts mood and realism.


Focus

Examples:

  • Soft focus

  • Bokeh

  • Deep focus

These adjust depth of field and sharpness.Using keyword controls reduces the need to overload the text prompt.

Last updated