Neural networks have stopped being just a toy. Today, it’s your personal stylist, artist, and producer all in one. We’ve gathered 5 magnificent prompts that will help turn ordinary photos into works of art, create unique stickers for messaging, and even “shoot” a movie without a camera.

The time for experimentation has come. Save these instructions and add magic to your feed!

1. Minimalist packaging shot.

 Prompt:

Minimalist packaging shot of a {Object}: the object rests on a light matte surface, captured mid-transformation into a volumetric pixelated form. One half remains intact, while the other smoothly disintegrates into large floating cubes scattering outward. Each cube clearly reveals the textures, ingredients, and colors of the original object. Soft studio lighting with realistic shadows, shallow depth of field, precise perspective and composition, hyper-detailed, geometric abstraction, subtle motion blur on the moving cubes, high resolution, cinematic close-up.

2. 3D Pixar-style character portrait with facial likeness preservation.

 Prompt:

Subject: 3D cartoon character portrait based on uploaded image (character reference: preserve facial features, head shape, and human proportions from photo). Enlarged head with playful exaggeration, expressive eyes and smile, professional 3D animation style.
Composition: Shoulder-length portrait, character facing camera directly, centered framing, rule of thirds for eye placement.
Action: Neutral relaxed pose, slight head tilt, engaging eye contact with viewer.
Setting: Minimalistic soft gradient background, muted warm grey to pale blue transition, no distracting elements.
Style: Smooth polished 3D render, clean materials with subtle subsurface scattering on skin, leather and fabric with neat plastic-like finish as in modern Pixar/Disney animation design. Octane render aesthetic, high-quality shading.
Lighting: Soft studio key light from camera left at 45°, gentle fill light from camera right, subtle rim light for subject separation. Even diffused illumination, no harsh shadows.
Color Palette: Warm skin tones, neutral background (hex #E8E8E8 to #D0D8E0 gradient), saturated but not oversaturated character colors.
Technical: Vertical 9:16 aspect ratio, 85mm portrait lens feel, shallow depth of field with soft background bokeh, subtle render grain for realism.
Constraints: Preserve facial recognition from uploaded image reference. Maintain original lighting and composition. No watermark, no extra text, no additional characters, no busy background patterns. Exaggeration limited to head size and expressiveness only — keep body proportions human.

 

3. Creating a style moodboard from your reference.

 Prompt:

[STYLE / MEDIUM]
Professional cinematic style guide layout, high-end visual development board, clean graphic design presentation. Photorealistic texture samples and accurate color reproduction.

[SUBJECT]
A vertical “Visual Style Reference Board” derived exclusively from the uploaded image. The board contains no new characters or scenes, only extracted visual elements arranged in a structured grid.
Components to include:

1. Color Palette: 5-6 distinct color swatches extracted from the image, labeled with their hex-code approximation (visual only).
2. Texture Macros: Extreme close-up squares showing material details (e.g., skin pores, fabric weave, metal rust, wet asphalt) found in the source.
3. Lighting & Atmosphere: Small rectangular fragments showing the specific lighting gradient, fog density, or volumetric rays from the original scene.
4. Composition Studies: Miniature cropped frames highlighting key framing techniques from the source.

[COMPOSITION]
Vertical layout. Organized into clear horizontal sections:

• Top: Title “CINEMATIC STYLE GUIDE” in bold, clean sans-serif typography.
• Upper Middle: Row of extracted Color Swatches.
• Middle: Grid of Texture Macro close-ups (2×2 or 3×2).
• Lower Middle: Strips of Lighting and Atmospheric references.
• Bottom: Environment detail studies and mood fragments.
  White or very light grey neutral background behind the board elements to ensure clarity. Clean spacing between elements.

[ACTION / POSE]
The layout presents a deconstructed analysis of the uploaded image. Elements are “pinned” or “placed” neatly as if on a digital storyboard. No motion; static presentation.

[ENVIRONMENT]
Neutral, clean studio backdrop for the board itself (light grey or white). The content of the board reflects the environment of the uploaded image (cyberpunk, noir, fantasy, etc.), but the container is a clean design space.

[LIGHTING]
Even, flat, diffused studio lighting illuminating the board itself to ensure colors and textures are visible without glare. The internal lighting of the texture/lighting samples must match the source image exactly.

[COLOR PALETTE]
Dominant colors extracted strictly from the uploaded image for the swatches. The board background is neutral (white/light grey) to make the extracted colors pop.

[TECHNICAL]
Shot on high-resolution digital display, sharp focus on all elements. 8k resolution feel. Vector-sharp text.

CRITICAL: Do NOT generate new characters, faces, or full scenes. Only extract and display fragments, swatches, and textures from the uploaded image.
CRITICAL: Maintain exact color accuracy from the source image for the palette swatches.
CRITICAL: Keep the layout organized and grid-aligned. No overlapping or messy collages.

4. Generating a new frame strictly according to the moodboard.

What it does: Takes the moodboard created above and uses it as a strict reference for generating a new scene with your desired character or object.

 Prompt:

Strictly adhere to the attached visual style reference: apply the same palette, lighting, atmosphere and cinematic grading. The color grading and atmosphere must exactly match the attached style reference.

Ensure the subject interacts naturally with the environment.

5. Extract the object.

 Prompt:

An ancient stone statue of a praying monk, covered in green moss, cracked. Glowing electric blue neon circuit lines are running through the cracks, illuminating the stone from within. Deep twilight, misty forest setting, high contrast, photorealistic, cinematic lighting, 8k.

How to Get the Best Results?

1. Quality of originals: The neural network is not a magician, it’s an artist. The clearer your photo, the better the result.
2. Experiment: Don’t be afraid to change details in prompts. Want a red dress instead of green? Just write it!
3. Tools: Midjourney, Stable Diffusion, or built-in generators in ChatGPT, Qwen, Gemini work best for image processing.
4. Preserve identity: If the face is important, use functions like --cref in Midjourney or special tools for personality preservation.

Create, get inspired, and remember: the only limitation is your imagination!