How to Write AI Video Prompts in NadouPro

To write effective AI video prompts in NadouPro, start with the result you want, attach suitable references, and describe each shot in a structured order: subject, scene, action, camera, lighting, style, and continuity notes. NadouPro supports Text to Video, Image to Video, First/Last Frame, Multi-reference, Lip Sync, and canvas-based node workflows, so a good prompt should tell the model what to generate and also make clear which image, video, audio, or text references should guide the result.

A simple NadouPro AI video prompt formula

When you want a clear, controllable first generation, use this structure:

Subject + identity details + environment + action + camera movement + shot size + lighting + mood + reference inputs + continuity requirements

For example, instead of writing only “a cinematic city scene,” a stronger prompt should state who or what appears, where the action takes place, what changes during the clip, how the camera moves, and which reference assets need to remain consistent.

Start with the right creation mode

Your prompt should match the NadouPro mode you are using. A pure text idea works differently from a prompt guided by a reference image, first frame, last frame, or motion video.

Goal	NadouPro mode to consider	What your prompt should emphasize
Create from an idea	Text to Video	Scene, subject, action, cinematic language, mood
Bring a still image to life	Image to Video	Motion, camera movement, what should stay unchanged
Control a transition	First/Last Frame	How the first frame changes into the last frame
Use multiple visual references	Multi-reference	Which reference controls the character, scene, costume, props, or style

Use references deliberately on the canvas

NadouPro’s canvas workflow lets creators combine text, image, audio, and video through nodes. In prompts, you can use @ references to point to specific assets, such as a character image, scene reference, or motion clip. This helps separate creative instructions from source materials.

One practical reference pattern is:

• Character reference: Use a character image or multi-view character setup when identity consistency matters.
• Scene reference: Use location images, panoramas, or visual boards when the space must remain recognizable.
• Motion reference: Use a video reference when body movement, rhythm, or performance matters.
• Audio reference: Use audio when the video depends on rhythm, sound, or Lip Sync.

Build prompts in layers

1. Define the subject

Clearly state who or what appears in the shot. Include stable details that should not drift, such as clothing, silhouette, age range, key props, or product form. Avoid putting too many conflicting identity cues into one sentence.

Use @character_ref as the main character. Keep the same face shape, hairstyle, clothing colors, and overall silhouette throughout the shot.

2. Define the space

Describe the location as a filmable space, not just a background. Mention the environment, depth, key light direction, important foreground or background elements, and atmosphere.

The scene is a quiet rooftop garden at night, with wet stone flooring, distant city lights, soft rim light from the left, and a calm, reflective mood.

3. Define the action beats

Video prompts need motion. Explain the beginning, middle, and end of the action. If the clip has multiple beats, number them so the model can follow the sequence more clearly.

Beat 1: The character looks down at the object in their hand. Beat 2: The character slowly turns toward the camera. Beat 3: The background lights grow brighter as the camera moves closer.

4. Add cinematic language

NadouPro supports cinematic shot language in prompts. When these terms help define the image, you can use terms such as Close-Up, Medium Shot, Wide Shot, Over-the-Shoulder Shot, Low Angle, High Angle, Dolly In, Pan Left, Tilt Up, Orbit, Tracking Follow, or Static Shot.

Medium Shot, slow Dolly In, Eye Level camera, shallow Depth of Field, soft Key Light and subtle Rim Light.

5. State continuity requirements

If the video is part of a sequence, add continuity instructions. These instructions are especially useful when using Image Fusion, Multi-reference, First/Last Frame, or connected canvas nodes.

Keep the same character identity, clothing, prop positions, scene layout, color palette, and time of day as @previous_shot.

Adaptable prompt templates

Text to Video template

Create a [shot size] video of [subject] in [location]. The action starts with [opening beat], then [main motion], and ends with [final pose or event]. Camera: [movement and angle]. Lighting: [key light, contrast, mood]. Style: [visual tone]. Keep the scene coherent and avoid adding unrelated objects.

Image to Video template

Animate @image_ref into a short video. Keep the subject, clothing, object shapes, and background layout consistent. Add [specific motion] while the camera [camera movement]. Preserve the original lighting direction and overall color mood.

Multi-reference template

Use @character_ref as the main character, @scene_ref as the environment, and @prop_ref as the object design. Generate a [shot size] scene where [action]. Camera [movement]. Keep the identity, scene geometry, and prop design consistent throughout the clip.

First/Last Frame template

Generate a transition from @first_frame to @last_frame. The motion should progress through [transformation or movement]. Keep the camera path smooth, preserve the main subject identity, and make the change feel physically continuous.

Motion Control template

Use @motion_video as the motion reference. Apply the same body movement and rhythm to @character_ref in [new scene]. Make the lighting match the new environment and keep the character identity stable.

When to use NadouPro tools beyond prompts

Prompts are only one part of the workflow. NadouPro also provides tools that help organize, guide, and refine generation:

• Canvas: Arrange text, image, video, and audio nodes into a visual workflow.
• Storyboard node: Convert script text into storyboard cards for shot planning.
• 3D Director: Plan staging before generation by using 3D scene snapshots and reference exports.
• Prompt Optimization: Use the optimization entry point in the generation dialog when you need a prompt closer to production level.
• Blockbuster Prompts: Search for reference styles in the Workbench production process when you need cinematic inspiration.
• Clip: Assemble generated assets into a final edit.

Common prompt mistakes to avoid

• Being too vague: “Make it cinematic” is less powerful than specifying the shot size, lighting, camera movement, and action.
• Including too many actions at once: Break complex events into numbered beats or separate shots.
• No continuity anchor: If a character, prop, or location must stay consistent, attach or reference that asset.
• Conflicting camera instructions: Avoid requesting multiple incompatible camera moves in one short clip.
• Ignoring the mode: A Text to Video prompt should describe the whole scene, while an Image to Video prompt should explain what changes relative to the existing image.

Recommended workflow

1. Write a one-sentence goal for the video.
2. Choose the NadouPro mode that fits the goal: Text to Video, Image to Video, Multi-reference, First/Last Frame.
3. Add the necessary references to the canvas.
4. Write the prompt in layers: subject, scene, action, camera, lighting, style, continuity.
5. Generate a short test.
6. Change only the weakest layer, such as motion, identity, lighting, or composition.
7. Use connected nodes and the Clip workflow to continue toward a full sequence.

Next steps

If you are new to this workflow, start with one character reference, one scene reference, and one clearly numbered action beat. Then use Storyboard, 3D Director references, and Clip assembly in NadouPro to expand into multi-shot planning.

Contains potentially incorrect AI-generated information. Please verify independently.

Contains AI-generated information that may be incorrect. Please verify independently.