Where to Find AI Prompt Examples? NadouPro Prompt Library: Image and Video Prompt Writing Breakdown

Key takeaways

• Image prompts and video prompts use two different writing methods: images focus on composition, light and shadow, and style, while videos also need action, duration, camera movement, and visual stability

• A reusable prompt becomes a template when broken down: subject, scene, style, action, constraints, and output format

• A weak prompt says "make a good-looking product video"; a strong prompt explains where the subject is, how it moves, how the camera moves, and what must not change

• Archive prompts by use case, save them together with the generated results, and next time you can simply replace the subject while reusing the structure

• NadouPro's Blockbuster Prompts feature includes many verified prompt structures that you can apply and modify directly

First decide whether you need an image or a video

Before writing a prompt, decide the output type first. This step determines the direction of the entire prompt.

• For posters, product images, portraits, concept art, or map illustrations, write an image prompt

• If you need action, camera movement, transitions, or a demonstration process, write a video prompt

• When character appearance, product details, or visual style need to stay consistent, attach a reference image

• Constraints do not need to cover everything; add limits only where mistakes are likely

The same subject can be made into either a static image or a video, but the prompt cannot be copied directly. Video adds a motion component that requires explicit instructions, while static images do not.

Image prompt examples

The following examples cover several common needs, including products, scenes, portraits, and illustrations. You can adapt them directly into your own version:

A studio-style product image of an insulated lunch box, natural morning light by a window, neutral background

A wide-angle view of a modern home office, walnut desk, soft shadows

A documentary-style portrait of an urban beekeeper holding a honeycomb frame, natural light

A top-down flat lay of travel items, clear labels, neatly arranged

A watercolor-style town map with simple landmark labels

Skincare products placed on a damp stone surface, realistic texture, soft reflections

A cozy reading corner, green armchair with a warm floor lamp

A minimalist poster for a seed exchange workshop, bold typography

A macro photography-style strawberry tart, glossy and translucent fruit surface

A set of 3D icons for a bookkeeping app, unified lighting

What these examples have in common is that each one clearly describes the subject, lighting, and style, without vague phrases such as "make it prettier" or "make it more premium."

Video prompt examples

Compared with image prompts, video prompts add three more components: action, camera, and stability:

The camera slowly pushes toward a steaming cup of coffee in a quiet morning kitchen

A fixed-camera rotating product display, white background, packaging text remains clear at all times

A city street after rain, neon reflections, slow walking pace

A teacher points to an animated diagram, clear annotations, clean classroom environment

A running shoe lands on a track and kicks up fine dust, low-angle shot

A time-lapse photography style shot of a seedling breaking through the soil and growing, soft daylight

A close-up of cooking steps, a hand sprinkling salt, no camera shake

A route animation on a map from the airport to the hotel, travel guide style

A floor plan gradually changes into a furnished layout, smooth top-view presentation

Historical year cards appear one after another, with map locations and icons

In NadouPro's canvas workflow, these prompts can be assigned directly to different Video node instances for batch generation, and then the best shots can be selected for the editing stage.

What does a complete video prompt with dialogue look like?

Beyond purely visual prompts, videos involving character dialogue and emotional changes need a more complete structure. Taking a typical scenario as an example, a prompt for a short beauty product livestream video could be written like this:

A beauty livestream room, warm yellow lighting illuminates the tabletop, with lipstick samples placed on both sides. A white beauty blogger raises a matte dusty-rose lipstick and says in a sweet, fresh voice: "A perfect pick for warm undertones! It brightens the complexion without drying out, with a soft-matte makeup look all day." Soft beauty-themed background music plays in the background.

Here is another example of a short everyday-life skit, where two characters appear in turn, each with lines and reactions:

A kitchen in the early morning, sunlight falls through the window onto the counter, and a fried egg sizzles in the pan. The boyfriend brings a burnt fried egg to the table, raises his eyebrows proudly, and says in a cheerful tone: "Try my breakfast made with love!" The girlfriend leans in to smell it, raises her eyebrows, and teases: "I can feel the love, but it is a little burnt." They look at each other and laugh together, saying: "It is a little burnt." The camera cuts from a close-up of the fried egg to the two of them smiling at each other.

The key to this kind of multi-character prompt is to make clear who is speaking in each segment, who stays silent, what the emotion is, and how the camera cuts. NadouPro's Screenwriting Assistant and Script Space can first organize the dialogue and scenes, then split each scene into a separate video prompt, which is more controllable than writing the entire plot all at once.

Break examples into templates

The real value of prompt examples lies in extracting reusable structures. The roles of each component in images and videos are as follows:

• Subject: in images, it is the main object or scene; in videos, it is the same, but you also need to add what it is doing

• Style: both need to specify the medium, lighting, and color tone

• Action: in images, it refers to pose or composition; in videos, it refers to motion and rhythm

• Control: in images, it means background and restrictions; in videos, it means camera movement and stability requirements

• Output: for images, specify aspect ratio or arrangement; for videos, specify duration and distribution platform

A product image template can be fixed into this structure:

[Product] placed on [surface], [lighting], [shot size], [mood], [background], [details that need to remain clear]

A video template is:

[Subject] [action], [camera movement], [scene], [duration], [style], [elements that remain still]

The value of a template is that every component has a clear responsibility: the subject tells the model what matters most, the scene provides context, the lighting determines realism, and constraints prevent the image from drifting off course. When the generated result is not ideal, revise the template instead of guessing blindly: if the frame is too crowded, simplify the scene; if the action is too large, reduce the range; if the style is wrong, describe the medium more specifically.

Side-by-side revision: the gap between weak and strong prompts

The fastest way to understand prompt writing is through comparison. A weak prompt is only one sentence: "make a cool product video," leaving the model to fill in everything by itself. A strong prompt would be written as:

A matte black water bottle slowly rotates on a white studio table, soft shadow, fixed camera, realistic reflections, the bottle label remains clear and readable at all times

A strong prompt usually contains five elements: a clear subject, visible action or composition, a scene that serves the purpose, a clear style or lighting setup, and restrictions that prevent unexpected changes. If any one of these is missing, the model will improvise in that area.

Prompts need to match the model's behavior

Different models respond differently to prompt length: some can handle very long descriptions, while others are more sensitive to short instructions. NadouPro integrates dozens of mainstream generation models, so the same prompt can be run several times with different models to compare results. If the prompt is ignored, simplify it; if the image drifts, add constraints. A good prompt is not the same as a long prompt. The point is to make the key details clear.

When you are unsure about a shot, you can also use NadouPro's Blockbuster Prompts feature directly. Its built-in mature prompt structures have been verified extensively, and choosing a similar one to modify the subject and scene is faster than writing from scratch.

Build your own prompt library

Manage prompts separately by output type: image prompts, video prompts, copywriting prompts, and editing prompts should each have their own place. Mixing them together slows down reuse. When saving, it is recommended to include four types of information:

• The original prompt

• The corresponding generated result

• The model and parameters used

• A one-sentence note on why this prompt worked

Tag them by use case, such as product image, explainer short video, poster, character, or opening hook, so they are easier to find. NadouPro's canvas space can keep prompts, assets, and generated results in the same project. During team collaboration, others can directly see what effect each prompt produced. It supports up to 100 people online at the same time, so the prompt library is naturally shared across the team.

Common questions

Q: What kind of prompt counts as a good prompt?

A good prompt is specific enough that someone can imagine the image before it is generated: subject, scene, style, lighting, action, aspect ratio, and constraints are all in place. The line between useful and vague lies in whether the model receives visible image instructions or only emotional adjectives.

Q: Is there a big difference between writing image prompts and video prompts?

Yes, the difference is significant. Image prompts revolve around subject, composition, medium, lighting, color, and use case; video prompts also need motion, rhythm, camera behavior, and what should remain stable from frame to frame. A static image can tolerate a certain degree of stylistic ambiguity, but video cannot, because every second of footage must connect with the next.

Q: How long should a prompt be?

Long enough to remove key uncertainty, but short enough to read smoothly. A simple image only needs one sentence, while a complex video, branded content, or character-driven plot may require several clauses. If every detail has a purpose, being a little longer is fine; if the details conflict with one another, it should be simplified.

Q: Should prompt examples be saved?

Yes. It is also recommended to save them together with generated results, parameter settings, reference images, and a note explaining why they worked. A personal prompt library saves the time of rewriting from scratch; what you reuse is the structure, not the sentence. Tagging by use cases such as product image, explainer short video, poster, character, and opening hook is the most practical approach.

Q: Why does the same prompt produce different results each time?

Generation models have inherent randomness, and parameter settings, model versions, and different interpretations of vague words can all cause variation. When consistency is needed, use clearer constraints, reference images, and a verified prompt structure. A prompt is a directional instruction, not a mechanical command that guarantees exactly the same output every time.

Explore more features on NadouPro: NadouPro

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