灵感图库
探索和使用精选的 AI 图像生成模板 (共 6,519 个)
90s-style film portrait from a reference photo
A long, detailed prompt to turn a reference photo of a man into a warm, grainy 1990s film-style portrait in a cozy indoor setting.
Where’s Waldo Star Wars scene
Generates a dense “Where’s Waldo” style illustration featuring all Star Wars characters on Tatooine, including Waldo.
Cat drinking cola illustration
A playful prompt that makes a cat drinking Coca-Cola, suitable for fun posts or memes.
Hieronymus Bosch-style AI future painting
Creates a surreal painting about the state of AI in 2025 in the style of Hieronymus Bosch.
Two-layer neural network diagram in Stephen Biesty style
Generates a diagram of a two-layer neural network illustrated in the detailed, cutaway style of Stephen Biesty.
Storyboard-style action-following frame
This prompt generates a new frame that follows the sketched annotations, continuing a stampede scene toward the camera.
Infographic from pasted text
This prompt turns a block of text, such as a summary, into a clean infographic layout.
Group image from multiple character references
A simple prompt that asks Nano Banana Pro to combine multiple individual character reference images into one group shot.
Convert scene to low-res pixel art
A brief prompt instructing the model to recreate an entire scene, including water splashes and background, as low resolution pixel art.
Photoreal K-pop selfie on rooftop
A very detailed JSON prompt for an 8K photoreal rooftop selfie of a K-pop idol named LUNA, including micro facial details, lighting, environment, and camera metadata.
High-angle fashion portrait prompt for Gemini 3 Pro
Reply contains a detailed prompt for a high-angle fashion portrait of a young Asian man on marble stairs.
Sci-fi cinematic triptych and expansions
A pair of prompts to create a cinematic sci-fi triptych in a 4:5 composition and then expand each panel into full 4:5 scenes, inspired by Roger Deakins and IMAX framing.
Cards, Shakespeare, and Nabokov scenes
Three short prompts: recolor trading cards to black, depict Shakespeare writing a famous line, and show Nabokov writing about aurochs and angels with butterflies around.
Cyberpunk woman in neon frame
A structured JSON prompt of a lone woman in a foggy cyberpunk city street, silhouetted in a glowing red rectangular frame with teal and red lighting.
Chess legends hanging out in a bar
A simple scene prompt asking for Bobby Fischer, Mikhail Tal, and Magnus Carlsen together in a bar, ideal for character group illustrations.
Elderly bearded man black-and-white portrait
A straightforward portrait prompt for a highly realistic black-and-white photo of an elderly bearded man with strong side lighting and detailed wrinkles.
Stable Diffusion educational anime diagram
A prompt to create a detailed diagram explaining Stable Diffusion with cute chibi anime teachers, aimed at high school-level education.
GARP biology for Tregs illustration
A concise biology prompt to draw an illustration explaining GARP biology for regulatory T cells, meant for scientific visualization.
Cheerful minimal edit to existing art
A very short edit command asking the model to make an artwork more cheerful with minimal changes, useful for subtle emotional adjustments.
Guts and Griffith on movie set
A pair of imaginative prompts: one placing Guts and Griffith on a movie set, and another surreal Mars food stall scene with Einstein, Sun Wukong, and Shakespeare.
Comic from single image
A prompt that turns one uploaded image into a multi-panel comic, ideal for narrative expansion and comic layout generation.
Formula graph on whiteboard
A short instruction-style prompt to draw a formula on a whiteboard and render its graph beneath, good for educational or technical illustrations.
Clothing items laid out on asphalt
A very short edit-style prompt asking the model to separate and place each piece of a person’s clothing on asphalt, useful for testing object extraction and rearrangement.
Infographic about past image model failures
A pair of prompts to create and refine an infographic explaining why AI models struggle with specific visual tasks, referencing clock times and training data issues.