What you'll make
A checked set of generated visuals matched to the source assets and intended format.
Create practical visuals with ChatGPT Images
Who this is for
Creators, marketers, founders, and small teams producing visual work from their own source assets.
Step by step
Run the workflow
- 01
Create an advertisement from a style reference
Use the source inputs to complete this stage and produce a result that can be checked before continuing.
- Click here to open ChatGPT (opens in a new tab)Official tool link added because the source does not include a working tool destination.
- Copy the prompt below and replace any project-specific subject, product, location, or format details.
- Attach the source assets in the same order described by the prompt, then run one controlled test.
Working assetCreate an advertisement from a style reference prompt 1. Create an Ad with Reference Image Use the selfie girl as the main subject. Create a vertical poster ad inspired by the reference poster style. Show the selfie girl holding a Pocari Sweat bottle with the framing of wide lens, low angle from the reference photo. Make it a trendy Japanese- style commercial poster with playful collage elements. Bright, eye-catching, stylish, and fun. ar 9x16
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. Common failure modes- The model changes the subject, framing, text, product, or background instead of editing only the requested element.
- The result looks polished at first glance but contains broken hands, unreadable text, warped objects, or inconsistent lighting.
- 02
Add live-action VFX to a portrait
Use the source inputs to complete this stage and produce a result that can be checked before continuing.
- Copy the prompt below and replace any project-specific subject, product, location, or format details.
- Attach the source assets in the same order described by the prompt, then run one controlled test.
Working assetAdd live-action VFX to a portrait prompt Use the uploaded person as the main subject. Create a realistic live-action VFX. He holding a sword instead of the phone. Do not change his face or body. He is still in the original location. Add a two goblins enemy surrounding him and a small transparent glowing game-style stats UI on top of the goblins, he is staring at one of the goblin, ar 9x16.
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. Common failure modes- The model changes the subject, framing, text, product, or background instead of editing only the requested element.
- The result looks polished at first glance but contains broken hands, unreadable text, warped objects, or inconsistent lighting.
- 03
Create a five-page manga
Use the source inputs to complete this stage and produce a result that can be checked before continuing.
- Copy the prompt below and replace any project-specific subject, product, location, or format details.
- Attach the source assets in the same order described by the prompt, then run one controlled test.
Working assetCreate a five-page manga prompt Create a 5-page manga using the attached reference photos of Blacpink Lisa and Gordon Ramsay. Keep their appearance recognizable and consistent, but turn them into anime versions of themselves. Show a short story of Lisa confidently teaching Gordon Ramsay how to make Thai food, Gordon cooking pad thai, Lisa tasting it and scolding him because it tastes bad, and Gordon reacting in shock while trying again. Keep the dialogue minimal and short, with only a few expressive speech bubbles. Make the page feel like a real manga with clear narrative flow, energetic panel composition, cinematic framing, expressive reactions, and polished black-and-white manga styling.
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. Common failure modes- The model changes the subject, framing, text, product, or background instead of editing only the requested element.
- The result looks polished at first glance but contains broken hands, unreadable text, warped objects, or inconsistent lighting.
- 04
Generate reverse camera angles
Use the source inputs to complete this stage and produce a result that can be checked before continuing.
- ChatGPT Images 2.0
- 4. Create New Scenes from Another Perspective
- A: Give me the angle from the barista perspective. ar 9:16 B: Give me the angle from the customer perspective. ar 9:16

Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. Common failure modes- The model changes the subject, framing, text, product, or background instead of editing only the requested element.
- The result looks polished at first glance but contains broken hands, unreadable text, warped objects, or inconsistent lighting.
- 05
Create a five-panel film storyboard
Use the source inputs to complete this stage and produce a result that can be checked before continuing.
- Copy the prompt below and replace any project-specific subject, product, location, or format details.
- Attach the source assets in the same order described by the prompt, then run one controlled test.
Working assetCreate a five-panel film storyboard prompt Create a 1-page colored film storyboard using the attached reference photos. Use the uploaded man as the main character and keep his face accurate and recognizable across all panels. Use the uploaded cat as his companion, a fluffy cream ragdoll cat with blue eyes and slightly darker ears, face, and paws. Show a 5-panel cinematic storyboard with clear visual continuity and a real storyboard layout, not a comic. Include short scene descriptions under each panel. Scene 1: establishing shot, the man arrives home. Scene 2: he notices the cat looking up at him. Scene 3: he kneels and reaches toward the cat. Scene 4: emotional close-up of him holding the cat. Scene 5: final wide cinematic shot of both together in a quiet room. Colored, cinematic, professional storyboard style with panel borders and shot-planning feel. ar 16x9
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. Common failure modes- The model changes the subject, framing, text, product, or background instead of editing only the requested element.
- The result looks polished at first glance but contains broken hands, unreadable text, warped objects, or inconsistent lighting.
- 06
Create a commercial storyboard grid
Use the source inputs to complete this stage and produce a result that can be checked before continuing.
- Copy the prompt below and replace any project-specific subject, product, location, or format details.
- Attach the source assets in the same order described by the prompt, then run one controlled test.
Working assetCreate a commercial storyboard grid prompt 6. Create Storyboard for Product Ads Use the uploaded male portrait as the main talent and the uploaded orange earbuds as the product reference. Create a cinematic 3x3 storyboard grid for a premium earbuds commercial. 9 frames in one clean 16:9 layout. Show a sequence of ad scenes highlighting comfort, portability, premium translucent design, noise cancelling, and powerful immersive sound. Make it feel sleek, high-end, modern, and aspirational, with dramatic movie-like lighting and strong visual energy. Include bass explosion visual moments, soundwave effects, and dynamic motion in some frames. Mix product close-ups, lifestyle shots, and cinematic hero shots of the model using the earbuds. Add a short visual description beneath each frame. Clean storyboard presentation, photorealistic, polished product advertising style, appealing composition, orange color accents, realistic details. ar 16:9
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. Common failure modes- The model changes the subject, framing, text, product, or background instead of editing only the requested element.
- The result looks polished at first glance but contains broken hands, unreadable text, warped objects, or inconsistent lighting.
- 07
Create a premium product photograph
Use the source inputs to complete this stage and produce a result that can be checked before continuing.
- Copy the prompt below and replace any project-specific subject, product, location, or format details.
- Attach the source assets in the same order described by the prompt, then run one controlled test.
Working assetCreate a premium product photograph prompt Create a premium commercial photo of the Morinaga BAKE Creamy Cheese box, remove the hand. Keep the packaging accurate and clearly visible. Make it feel like a polished ad campaign visual with photorealistic lighting, strong styling, clean art direction, shallow depth of field, and luxury brand photography quality. Refine the background into a cleaner, more aesthetic Japanese snack store setting with soft blur, while keeping the product as the main focus. Crisp details, appealing composition, premium food product advertising look. ar 4:5
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. Common failure modes- The model changes the subject, framing, text, product, or background instead of editing only the requested element.
- The result looks polished at first glance but contains broken hands, unreadable text, warped objects, or inconsistent lighting.
- 08
Create a four-person group photograph
Use the source inputs to complete this stage and produce a result that can be checked before continuing.
- Copy the prompt below and replace any project-specific subject, product, location, or format details.
- Attach the source assets in the same order described by the prompt, then run one controlled test.
Working assetCreate a four-person group photograph prompt Create a realistic group photo of the 4 men with natural posing, and cohesive body proportions. Make the people look like they were photographed together in the same place at the same time. Clean composition, natural expressions. They are hanging out in singlets and home clothes like buddies in a remote village, one of them is holding a lollipop another one holding a Sparklers-Handheld. ar 16x9
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. Common failure modes- The model changes the subject, framing, text, product, or background instead of editing only the requested element.
- The result looks polished at first glance but contains broken hands, unreadable text, warped objects, or inconsistent lighting.
- 09
Create a natural group photograph
Use the source inputs to complete this stage and produce a result that can be checked before continuing.
- Copy the prompt below and replace any project-specific subject, product, location, or format details.
- Attach the source assets in the same order described by the prompt, then run one controlled test.
Working assetCreate a natural group photograph prompt Create a realistic group photo with natural posing, and cohesive body proportions. Make the people look like they were photographed together in the same place at the same time. Clean composition, natural expressions. They are hanging out in singlets and home clothes like buddies in a remote village. ar 16x9
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. Common failure modes- The model changes the subject, framing, text, product, or background instead of editing only the requested element.
- The result looks polished at first glance but contains broken hands, unreadable text, warped objects, or inconsistent lighting.
- 10
Create a YouTube thumbnail
Use the source inputs to complete this stage and produce a result that can be checked before continuing.
- Copy the prompt below and replace any project-specific subject, product, location, or format details.
- Attach the source assets in the same order described by the prompt, then run one controlled test.
Working assetCreate a YouTube thumbnail prompt Use the uploaded portrait as the main subject. Create a YouTube thumbnail with the same person in a close-up crop, shocked and surprised expression, reacting to a major AI news update. Add a bold headline that says "NEW UPDATE ON CHATGPT!" Clean and clickbaity thumbnail design, high contrast, modern tech background with subtle AI interface elements, green glow accents, polished layout, eye-catching and professional. ar 16:9
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. Common failure modes- The model changes the subject, framing, text, product, or background instead of editing only the requested element.
- The result looks polished at first glance but contains broken hands, unreadable text, warped objects, or inconsistent lighting.
- 11
Upscale an image to a cinematic 4K look
Use the source inputs to complete this stage and produce a result that can be checked before continuing.
- Copy the prompt below and replace any project-specific subject, product, location, or format details.
- Attach the source assets in the same order described by the prompt, then run one controlled test.
Working assetUpscale an image to a cinematic 4K look prompt Enhance the uploaded image into a pristine 4K cinematic version while preserving the subject with maximum fidelity. Keep identity, composition, key details, and original visual intent intact. Improve sharpness, texture, lighting quality, depth, and overall polish without changing the person or scene unnecessarily. Ultra-clean, high-definition, photorealistic result.
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. Common failure modes- The model changes the subject, framing, text, product, or background instead of editing only the requested element.
- The result looks polished at first glance but contains broken hands, unreadable text, warped objects, or inconsistent lighting.
- 12
Create a six-look fashion collage
Use the source inputs to complete this stage and produce a result that can be checked before continuing.
- Copy the prompt below and replace any project-specific subject, product, location, or format details.
- Attach the source assets in the same order described by the prompt, then run one controlled test.
Working assetCreate a six-look fashion collage prompt Use the uploaded man as the main subject and the outfit board as style reference. Create a fashion editorial collage of the same man in 6 distinct smart casual outfits. Keep his face consistent in every look. Show full body styling with neutral minimalist menswear, inspired by the reference color combinations. Arrange the 6 looks in a clean balanced layout on a cream studio background, no heavy borders. Add small handwritten style notes pointing to key clothing pieces. Modern, stylish, polished, soft studio lighting, premium menswear editorial look. ar 4:5
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. Common failure modes- The model changes the subject, framing, text, product, or background instead of editing only the requested element.
- The result looks polished at first glance but contains broken hands, unreadable text, warped objects, or inconsistent lighting.
- 13
Create a warm food advertisement
Use the source inputs to complete this stage and produce a result that can be checked before continuing.
- Copy the prompt below and replace any project-specific subject, product, location, or format details.
- Attach the source assets in the same order described by the prompt, then run one controlled test.
Working assetCreate a warm food advertisement prompt 12. Create Home Ad with Your Own Food Use the uploaded chocolate-coffee butter bun as the hero food item. Create a warm home- style advertisement in a cozy natural home setting, with beautiful plating, soft inviting lighting, and an authentic lifestyle atmosphere. Make the bun look fresh, rich, and appetizing, with clear texture and a homemade premium feel. Clean composition, photorealistic, commercially appealing food photography, like a premium food campaign with a homemade touch. ar 4:5
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. Common failure modes- The model changes the subject, framing, text, product, or background instead of editing only the requested element.
- The result looks polished at first glance but contains broken hands, unreadable text, warped objects, or inconsistent lighting.
- 14
Redesign a room in a wabi-sabi style
Use the source inputs to complete this stage and produce a result that can be checked before continuing.
- Copy the prompt below and replace any project-specific subject, product, location, or format details.
- Attach the source assets in the same order described by the prompt, then run one controlled test.
Working assetRedesign a room in a wabi-sabi style prompt Use the uploaded living room and dining area as the base space. Redecorate the interior in a warm wabi-sabi style while keeping the same room layout, windows, and overall architecture. Replace the current furniture and decor with a more refined natural look using light wood, linen, textured plaster, stone, earthy beige tones, soft brown accents, organic shapes, minimal low furniture, handmade ceramics, and subtle greenery. Make the space feel calm, elegant, cozy, and uncluttered, with soft natural lighting and premium photorealistic interior design mockup quality. ar 9:16
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. Common failure modes- The model changes the subject, framing, text, product, or background instead of editing only the requested element.
- The result looks polished at first glance but contains broken hands, unreadable text, warped objects, or inconsistent lighting.
- 15
Turn a floor plan into a top-down render
Use the source inputs to complete this stage and produce a result that can be checked before continuing.
- Copy the prompt below and replace any project-specific subject, product, location, or format details.
- Attach the source assets in the same order described by the prompt, then run one controlled test.
Working assetTurn a floor plan into a top-down render prompt 14. Create Interior Design from Floor Plan Top Down Use the uploaded floor plan as reference. Turn it into a realistic top-down interior render that shows how the home would look in real life. Keep the same room layout and proportions, including the living room, bedroom, bathroom, kitchen, and dining area. Redesign the whole space in a warm wabi-sabi style with light wood, limewash or textured plaster walls, stone, linen, soft beige and earthy tones, minimal furniture, organic shapes, handmade decor, and a calm uncluttered feel. Photorealistic bird's-eye view, clean styling, soft natural lighting, premium interior design visualization. ar 1:1
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. Common failure modes- The model changes the subject, framing, text, product, or background instead of editing only the requested element.
- The result looks polished at first glance but contains broken hands, unreadable text, warped objects, or inconsistent lighting.
- 16
Create an exploded-view food diagram
Use the source inputs to complete this stage and produce a result that can be checked before continuing.
- Copy the prompt below and replace any project-specific subject, product, location, or format details.
- Attach the source assets in the same order described by the prompt, then run one controlled test.
Working assetCreate an exploded-view food diagram prompt Use the uploaded burger image as reference. Turn the burger into a clean exploded-view component breakdown, separating each layer clearly in vertical order: top bun, sauce, bacon, melted cheese, beef patty, tomato, lettuce, bottom sauce, and bottom bun. Label each component clearly with neat callout lines. Remove the social media UI and turn it into a polished product diagram on a clean studio background. Modern educational layout, balanced composition, premium food infographic style, photorealistic, crisp details. ar 4:5
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. Common failure modes- The model changes the subject, framing, text, product, or background instead of editing only the requested element.
- The result looks polished at first glance but contains broken hands, unreadable text, warped objects, or inconsistent lighting.
- 17
Create a presentation deck preview
Use the source inputs to complete this stage and produce a result that can be checked before continuing.
- Copy the prompt below and replace any project-specific subject, product, location, or format details.
- Attach the source assets in the same order described by the prompt, then run one controlled test.
Working assetCreate a presentation deck preview prompt Turn the uploaded information into a premium presentation deck preview. Show a polished set of modern slides with strong layout hierarchy, clean typography, editorial spacing, and a high-end business design aesthetic. Include a title slide, insight slides, and a bar graph slide based on the data table. Make it visually impressive, clean, and presentation-ready, like a top-tier strategy or pitch deck.
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. Common failure modes- The model changes the subject, framing, text, product, or background instead of editing only the requested element.
- The result looks polished at first glance but contains broken hands, unreadable text, warped objects, or inconsistent lighting.
- 18
Create a brand deck preview
Use the source inputs to complete this stage and produce a result that can be checked before continuing.
- Copy the prompt below and replace any project-specific subject, product, location, or format details.
- Attach the source assets in the same order described by the prompt, then run one controlled test.
Working assetCreate a brand deck preview prompt Create a premium brand deck preview for a modern business. Include sample slides for brand overview, mission, positioning, audience, visual direction, color palette, typography, imagery style, and brand (Premium cinematic brand hero image for . Dark mode composition on Deep Midnight background (#071521) with frosted glass surfaces and soft luminous bloom. Centered hero emblem: concentric glowing orb in frosted aqua (#CFE9EA), soft teal (#8FD6D0), prism lavender (#B7A8D6), and cool sky cyan (#79C7E8), with a bright translucent center and cool prism edge light. Below the orb, wordmark in Space Grotesk Bold caps with subtle chrome-spectrum gradient. Tight letter spacing. Frosted aqua subtitle in Inter SemiBold. Monospaced metadata tags in JetBrains Mono in small caps along the edges. Generous negative space, 8 px grid rhythm, card radius 12 to 16 px. Mood: operator and educator, calm confidence, systems credibility, cinematic tech depth. Avoid neon, chrome heavy sci fi, retro Y2K, warm tones, busy textures, generic AI aesthetic.)
Use this source result as the visual comparison point for the step. Common failure modes- The model changes the subject, framing, text, product, or background instead of editing only the requested element.
- The result looks polished at first glance but contains broken hands, unreadable text, warped objects, or inconsistent lighting.
- 19
Create a business-card mockup
Use the source inputs to complete this stage and produce a result that can be checked before continuing.
- Copy the prompt below and replace any project-specific subject, product, location, or format details.
- Attach the source assets in the same order described by the prompt, then run one controlled test.
Working assetCreate a business-card mockup prompt Create a realistic premium business card mockup shown in a polished brand presentation style. Include both front and back card designs, elegant typography, refined spacing, and luxury paper or emboss-style finishing and the profile picture. Clean background, realistic shadows, professional mockup photography, modern premium brand identity look. Name "SIA"
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. Common failure modes- The model changes the subject, framing, text, product, or background instead of editing only the requested element.
- The result looks polished at first glance but contains broken hands, unreadable text, warped objects, or inconsistent lighting.
- 20
Create a logo design system
Use the source inputs to complete this stage and produce a result that can be checked before continuing.
- Copy the prompt below and replace any project-specific subject, product, location, or format details.
- Attach the source assets in the same order described by the prompt, then run one controlled test.
Working assetCreate a logo design system prompt Create a clean, modern, versatile logo design system for a brand. Show the main logo, symbol or icon variation, monochrome version, and practical mockup applications for website and source use. Make it feel professional, distinctive, scalable, and suitable for a strong modern brand identity. Minimal but memorable.
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. Common failure modes- The model changes the subject, framing, text, product, or background instead of editing only the requested element.
- The result looks polished at first glance but contains broken hands, unreadable text, warped objects, or inconsistent lighting.
- 21
Create campaign-ready marketing material
Use the source inputs to complete this stage and produce a result that can be checked before continuing.
- Copy the prompt below and replace any project-specific subject, product, location, or format details.
- Attach the source assets in the same order described by the prompt, then run one controlled test.
Working assetCreate campaign-ready marketing material prompt Create polished marketing material for this product or brand. Use a premium campaign- style layout with strong visuals, clear message hierarchy, modern branding, and high visual impact. Make it feel professional, commercial, and ready for social, print, or digital promotion.
Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. Common failure modes- The model changes the subject, framing, text, product, or background instead of editing only the requested element.
- The result looks polished at first glance but contains broken hands, unreadable text, warped objects, or inconsistent lighting.
- 22
Create a translated image
Use the source inputs to complete this stage and produce a result that can be checked before continuing.
- ChatGPT Images 2.0
- 21. Create a Translation Image
- Create a clean english translation visual that presents the translated original language.

Use this source result as the visual comparison point for the step. 
Use this source result as the visual comparison point for the step. Common failure modes- The model changes the subject, framing, text, product, or background instead of editing only the requested element.
- The result looks polished at first glance but contains broken hands, unreadable text, warped objects, or inconsistent lighting.
Before you use it
Final checks
- Check the full result for changed people, products, text, timing, or layout.
- Confirm the tool, model, controls, and cost before generating.
- Keep private or unreleased material out of third-party tools.
Share your work
Made it? Show me.
Post your finished result on Instagram and tag @terencesia. I'll share a few standout results with the community.
Tag @terencesia on Instagram (opens in a new tab)Next step
Run one low-cost test, record the first visible failure, and revise only the instruction or source asset tied to that problem.
Back to the tutorial library