Prompts that actually work
Every template here has been benchmarked against the GPT Image 2 model — the prompt engineering is already tuned.
54 curated, production-ready prompts that run on OpenAI's GPT Image 2 model. Copy any template, customize the variables, and generate in seconds — no prompt-engineering experience needed.
Click any prompt to open it in the editor. Free signup unlocks customization + generation.
Takes a user-uploaded portrait photo along with optional inputs (gender/age, hair texture, style preference, goals) and outputs a poster-quality hairstyle diagnosis report. The report includes geometric facial analysis overlays, 12 scientifically justified hairstyle recommendations with fit scores and rationale, a Top 3 conclusion, and a maintenance guide — styled for salon display, social sharing, or image consulting.
Takes 5 character parameters — character name, outfit style, primary color, shoe style, and color palette — and generates a professional ancient Chinese fantasy costume design reference sheet (角色设定图) rendered by GPT Image 2. Outputs a dark navy 3:2 widescreen infographic with 7 structured regions: identity block, full-body illustration, exploded outfit layers, accessory close-ups, shoe detail panel, three-view silhouettes, and color swatches. Ideal for game art pipelines, cosplay design, and fantasy character development.
Takes any theme, keyword, or short phrase as input and outputs a detailed GPT image generation prompt for a collector-grade LEGO brick-aesthetic typography concept poster. Automatically extracts core keywords, selects theme-aligned brick color palettes, derives visual metaphors, and composes multi-zone exhibition-style poster layouts. Ideal for designers, educators, brand creatives, and LEGO enthusiasts wanting typographic art concepts.
Takes a reference character image and a target year as input, then generates a single print-ready A4 portrait anime-style calendar poster. The output is a 3×4 grid showing all 12 months with accurate date grids, the character in 12 unique seasonal outfits and poses, watercolor pastel backgrounds, and hand-drawn doodle decorations. Ideal for fan art creators, kawaii merchandise designers, and personalized gift printing.
Transforms any city's metadata into a museum-quality vintage collectible postage stamp poster at 3:4 aspect ratio. Takes 8 user-supplied variables (city name, native script, air mail translation, denomination, issue year, coordinates/tagline, country, and stamp date) and outputs a richly detailed, anatomically authentic philatelic art poster in a 1950s–1970s lithograph aesthetic with perforated edges, postmark overlay, and retro travel-poster illustration.
Takes five brand identity variables (brand name, tagline, product type, accent color, hero voice-over line) and generates a single print-ready image via GPT Image 2 — a 12-panel cinematic storyboard arranged with labeled still-frame scenes, warm cream palette, 35mm film-grade aesthetics, and a branded footer. Ideal for luxury fragrance campaigns, editorial pitches, and high-end lifestyle brand presentations.
Recebe descrições textuais e parâmetros visuais (tema, estilo, iluminação, composição, formato) e gera prompts fotorrealistas e cinematográficos em inglês, otimizados para ferramentas de IA como GPT Image, Runway, Kling e Sora. Ideal para criadores de conteúdo, designers e profissionais de marketing que precisam de imagens e vídeos indistinguíveis da realidade.
Takes personal diary content (text narration, dated entries) and optional reference photos as input, then outputs structured comic panel sequences with cohesive visual generation prompts. Automatically segments diary events into panels, rewrites content in comic-style language, applies consistent character styling, and synthesizes ready-to-use image generation prompts. Ideal for turning daily journals into stylish illustrated comic diary pages.
Describe your fight scene — combatants, elemental powers, and mood — and get a director-ready 8-panel stick-figure storyboard prompt. Each frame includes motion trails, elemental FX, and cinematic camera directions. Perfect for manga artists, animators, and game designers who want to previs action sequences fast.
Takes subject photos, descriptions, scene details, and camera settings as inputs to generate hyper-detailed image. Recreates early 2000s/2010s consumer digital camera aesthetics with nested LCD-screen compositions, period-accurate UI overlays (timestamps, battery icons, mode indicators), and candid portrait styling. Ideal for nostalgic photography art, retro-style portrait generation, and era-accurate digital camera simulations.
Takes a reference photo and 6 user-defined variables (easel style, outdoor setting, outfit, caricature art style, and optional direction) to generate a single photorealistic image. The output portrays the subject sitting beside an easel displaying a recognizable caricature of themselves. Ideal for portrait artists, social media content creators, and editorial photography concepts.
Takes an uploaded anime character illustration plus dance style parameters and outputs a fully annotated, production-ready dance reference sheet prompt. Generates structured multi-panel layouts in Japanese anime design document style, including turnaround views, 6-step pose sequences, expression charts, hand gestures, footwork diagrams, and rhythm memos. Ideal for animators, idol game developers, and character designers.
Takes an uploaded character reference image and analyzes visible costume components, colors, aterials, and style. Outputs a structured multi-view costume breakdown sheet prompt ready for GPT Image generation, plus a reusable IP-neutral style template. Ideal for costume designers, concept artists, and game/anime creatives needing detailed garment reference sheets.
Takes a product name, optional product image (any quality), optional copywriting, and style/audience preferences as input. Outputs a publication-ready horizontal commercial banner image alongside a creative rationale and the raw generation prompt used. Ideal for e-commerce ads, social media banners, promotional campaigns, and retail display materials requiring agency-quality visuals without a design team.
Recebe tema educativo, região brasileira, descrição de cena e personagens ativos e gera storyboards completos de histórias em quadrinhos infantis. Produz prompts visuais em inglês para geração de imagem no estilo cartoon brasileiro (referência Turma da Mônica), diálogos em português do Brasil e notas pedagógicas integradas. Ideal para educadores, autores de livros infantis e criadores de conteúdo cultural brasileiro.
Takes a character group reference (image or text description) and generates a vivid, natural-language image generation prompt featuring all characters in a dynamic 'glancing back at the viewer' ensemble portrait composition. Supports 2–12 characters, multiple art styles, and customizable mood, setting, and color tone. Ideal for fan art, franchise portraits, family scenes, and team illustrations.
Takes a user-provided theme (e.g., cat, Mario, strawberry) along with style tone, color count, and complexity level, then outputs a high-resolution pixel-art style Perler/Hama bead pattern image. Features a single centered subject on a pure white background using 6–12 flat solid colors. Ideal for beginner and intermediate crafters looking to replicate patterns with physical Perler or fuse beads.
Transforms any city name into a premium editorial-style illustrated typography travel poster. Takes six inputs — city name, color palette, background color, tagline, descriptor words, and subtitle. Each letter of the city name becomes an illustrated window of local landmarks and culture, arranged in a 16:9 widescreen five-zone layout. Ideal for rapid multi-city poster generation with zero post-editing.
Takes an uploaded food photo plus brand and menu details, then outputs a ready-to-render GPT-Image prompt for a 2:3 vertical Food Hero Explainer Poster. Preserves the dish's original appearance while elevating it to premium commercial quality with brand zone, ingredients panel, selling points, price block, and social CTA row. Ideal for restaurant ads, delivery platforms, and social media campaigns.
Takes a user-uploaded themed Bingo card image and produces one shuffled variant by repositioning existing cell elements only. It extracts all cell contents (text, icons, colors, fonts), randomly reassigns their grid positions using a Fisher-Yates shuffle, and reconstructs a visually faithful output card. The FREE cell stays locked. No content is added, removed, or visually altered. Ideal for game hosts needing unique card variants from a master template.
接收用户输入的修仙世界内容变量(报纸名称、主新闻标题、主视觉人物、第二新闻主题、萌宠栏目、底部广告),自动生成信息密集、多栏网格排版的竖版古风修仙报纸海报提示词。适用于AI绘图平台生成古代报纸风格的修仙幽默海报图像。
Takes a character reference image and styling preferences as input, then outputs detailed image-generation prompts for multiple outfit variants. Automatically extracts character traits (skin tone, hair, body type) from the uploaded image and locks them as a consistency anchor across all generated looks. Ideal for character designers, AI artists, game developers, and fashion concept creators who need structured, reusable outfit prompt templates.
Takes tactical FPS map parameters (map name, theme, view angle, key zones, art style, atmosphere, and special elements) and outputs a single fluent natural-language image generation prompt. Designed for CrossFire and Counter-Strike style map concept art, covering archetypes like desert, urban, industrial, and castle environments for use with AI image generators.
Takes game parameters (type, theme, grid size, content type, card count, color mode) and generates a structured, print-ready game card image prompt. Supports Bingo, Word Bingo, Trivia Grids, and holiday variants. Outputs a natural-language image generation prompt optimized for GPT Image, with randomized cell content, column constraints, FREE space handling, and black-and-white print formatting for home printing on A4/Letter paper.
Takes user-provided male character details, Chinese fantasy scene elements, calligraphy text, and style preferences as inputs. Outputs an ultra-detailed English image generation prompt structured around a double exposure silhouette poster composition, blending ink wash aesthetics, xianxia architecture, eastern dragons, and anime art styles in a teal-green monochromatic mood. Ideal for game promo art, fantasy poster design, and Genshin-style fan art creation.
Takes a franchise character name and series as input, then automatically derives personality archetype, signature colors, pose style, and outfit redesign parameters to output a single cohesive natural-language image generation prompt. Produces vertical anime × streetwear × graphic poster prompts ready for GPT Image, ideal for fan art, creative campaigns, and character reimagining projects.
Takes a character name, visual description, and optional reference image as inputs, then generates a detailed GPT Image prompt for a collector-grade luxury anime commemorative stamp sheet featuring 6–9 emotionally distinct stamps. Each stamp includes gold foil borders, perforated edges, wax seals, postmarks, and serial numbers, with strict full-face visibility enforced across all stamps. Ideal for fan art collectors, anime merchandise designers, and philatelic-themed illustration projects.
Takes six broadcast scene parameters (sport type, match time, outfit, team colors, scoreboard text, and network logo) and outputs a photorealistic GPT Image 2 prompt plus a 16:9 freeze-frame still of a glamorous female spectator in a packed stadium crowd. Ideal for viral social content, cinematic sports edits, and AI video seed frames.
Takes a character reference image and optional styling hints to generate a fully scripted 16-frame 'Get Ready With Me — Outfit Edition' storyboard collage prompt. Outputs a high-fidelity image generation prompt for a 4×4 grid at 1080×1350px, along with detected visual style and character description. Ideal for content creators, illustrators, and fashion storyboard artists.
Takes a character's name and personality traits as input and outputs a single anime-style illustration of that character proudly holding an open sketchbook displaying their own hilariously idealized, personality-driven self-portrait. Useful for fan art creation, character visualization, creative writing companions, and anime-style OC (original character) showcases.
Takes user-defined character features (hair, eyes, outfit, accessories, background, mood, color palette, typography) and outputs a production-ready Midjourney natural-language prompt for psychedelic anime album cover art. Ideal for creating Y2K retro-futurist Vocaloid/J-pop idol posters with maximalist superflat graphic design aesthetics and neon rave visuals.
Input a character concept with optional details (hair, eyes, outfit, accessories, pose, background) and receive a fully structured Midjourney/GPT Image prompt for a full-body kawaii streetwear anime character illustration, suitable for game or commercial art use.
Transform a person's description or photo reference into a cinematic 3D scene where the subject dramatically breaks through a torn paper screen, surrounded by a realistic social media profile UI.
Takes a brand name (well-known or custom) along with mascot visuals, brand colors, and slogans, then outputs a hyper-detailed anime hypebeast streetwear commercial poster in a vertical Japanese ad layout. Ideal for marketers, streetwear brands, and creative agencies wanting culturally resonant, visually dominant brand poster concepts.
Takes an uploaded anime character reference image and optional style tone preference, then analyzes face shape, hair type, color, and aesthetic to generate 5 best hairstyles, 5 hairstyles to avoid, and 5 color palettes. Ideal for fan artists, character designers, and anime enthusiasts.
Input a Langrisser II character name with optional faction, platform version, and layout preferences. Outputs structured image generation prompts for high-definition final-class portraits and class promotion tree diagrams, formatted as encyclopedia-style visual atlas entries faithful to the 1994 JRPG source art.
Takes character customization inputs (hair color, eye color, outfit color, slime color, and optional extra details) and outputs a fully structured Midjourney-ready prompt string. Designed for generating deadpan chibi vinyl figure poster illustrations with streetwear, slime drip, graffiti typography, and toxic kawaii aesthetics. Ideal for designer toy artists, streetwear creators, and urban kawaii illustrators.
Upload your photo to receive a magazine-grade Before & After style upgrade report. The AI analyzes your look and generates a polished layout with upgraded outfit, layering, accessories, and styling — preserving your identity while delivering a visible, trend-forward transformation styled as a professional fashion stylist proposal board.
Takes a food concept (e.g., Malatang, Ramen, Bubble Tea) and optional parameters like gender, color palette, and cultural style, then outputs a fully structured, game-quality anime character design sheet prompt. Designed for use with GPT-image models, it generates multi-panel reference layouts covering turnaround views, expression sheets, outfit details, and color swatches for food personification characters.
Input any subject, concept, or knowledge point (with optional reference images and icon hints), and get a detailed English prompt for AI to generate a healing, 2D Japanese-style scrapbook infographic. Ideal for educators, designers, or knowledge visualization tasks.
Input a restaurant brand name and business category (plus optional details like slogan, color, and style) to generate a professional English image prompt for a multi-touchpoint brand visual matrix board, along with an inferred brand profile summary. Ideal for restaurant branding, packaging design, and visual identity projects.
Input a healing stationery brand name or series theme; outputs a detailed English flat lay mockup image prompt plus layout description. Ideal for cultural creative merchandise sets (20+ products) with dreamy forest/moon/animal aesthetics, ready for AI image generation tools.
Input a Chinese character and style preferences (font style, grid type, paper texture) to generate a print-ready 5-cell calligraphy practice template image prompt. Cell 1 shows the reference character, cell 2 shows a ghost tracing guide, and cells 3–5 are blank practice spaces.
Input a notable person's name and optional style preferences; output a detailed, editorial-grade infographic poster image prompt featuring a photorealistic/illustrative central portrait, career timelines, product references, annotated callouts, and bold graphic design elements. Ideal for tribute posters, educational visuals, and editorial design.
Input a brand name, campaign theme, and plush character name to generate a complete, advertising-grade image generation prompt for a commercial campaign poster featuring a plush/stuffed character. Suitable for brand marketing, seasonal campaigns, and IP character promotions.
Input a sacred beast theme (Azure Dragon/White Tiger/Vermilion Bird/Black Tortoise/Qilin) and visual state (activated/static) to generate a production-ready image prompt for a traditional Chinese Five Elements card border with protected center blank zone.
Input city, location, museum name, designer, style preference, and color scheme to generate a complete A1 exhibition panel prompt for a High-Tech/Modernist public art museum, including a bird's-eye rendering and full architectural analysis diagrams at 3 PM lighting.
Input a city name, its iconic landmarks, and a promotional slogan to generate a production-ready image prompt for a city poster. Output includes a detailed English prompt featuring S-curve silk composition, double-exposure scroll transition and ink-wash landmark illustration — optimized for AI image generation.
Upload any everyday photo and receive a cozy, journal-style annotated image prompt. The system detects visible elements, generates warm handwritten English captions (≤6 words each), maps semantic decorations (steam, stars, hearts, arrows), and assembles a complete image generation prompt that overlays thin white one-stroke outlines and dashed arrow guides over the original photograph.
Upload a portrait photo and provide persona details to generate a photorealistic LinkedIn profile card UI mockup prompt. Outputs a full image generation prompt preserving the character's facial likeness within a circular avatar frame, complete with LinkedIn UI elements (name, title, skills, endorsements, connections) in the official blue-white color scheme.
Input any domain (DevOps, UI Design, Data Science…) and get a prompt for a dense grid poster of real, brand-accurate tool icons on a 1:1/3:2/2:3 canvas. GPT Image 2 fits as many verified tools as the frame allows.
Input a game concept with theme, component type, paper size, and age group to generate print-ready tabletop game component images (boards, cards, tokens, score sheets) optimized for home printing with cut guides and CMYK-safe colors.
Every template here has been benchmarked against the GPT Image 2 model — the prompt engineering is already tuned.
Pre-configured aspect ratios, styles, and parameters. Edit one variable and hit Generate.
Download the original full-resolution image. No watermark, commercial use OK.
Browse the library above and click the one closest to what you want to generate.
Change the subject, style, or composition. The rest of the prompt stays optimized.
GPT Image 2 returns your image in ~10 seconds. Download, share, or iterate.
ChatGPT Image 2 (also called GPT Image 2) is OpenAI's image generation model accessible through the ChatGPT API. It excels at typography, complex composition, and understanding longer prompts compared to earlier image models.
No. Camika provides access to GPT Image 2 directly — sign in with email or Google to try it. Each generation costs a few credits.
Yes. Images you generate on Camika are yours to use commercially per OpenAI's content policy. No watermark is added.
Every template here is tuned against the GPT Image 2 model specifically. The prompt structure, variable placement, and aspect ratio defaults are picked from real production runs — not copy-paste guesses.
Yes. After signing in you can fork any template, modify it, and save it as your own. Your custom prompts can stay private or be published to the community library.
GPT Image 2 reads long, detailed prompts more reliably and renders text inside images much better. For comic panels with dialogue, UI mockups, or instructional graphics it's typically the strongest choice.
Sign up free. No card required.
Create free account