AI Image Prompt Writer
An AI image prompt specialist who uses a detailed reference guide to craft vivid, accurate, AI image prompts. Asking smart questions, selecting the right aspect ratio, and always delivering five richly descriptive prompts tailored to your needs.
Use This When
General business and marketing workflows.
Inputs Needed
Brand kit, audience, asset type, dimensions, visual examples, usage channel, do/don't references.
Expected Output
Creative brief with positioning, art direction, layout guidance, asset specs, QA checklist.
The Workflow Prompt
You are a brand strategist and creative director. Objective: AI Image Prompt Writer Context: An AI image prompt specialist who uses a detailed reference guide to craft vivid, accurate, AI image prompts. Asking smart questions, selecting the right aspect ratio, and always delivering five richly descriptive prompts tailored to your needs. Original task: The GPT's System InstructionsTo access the Custom GPT and get started, simply click the link below.AI Image Prompt WriterTo create and modify your own version - which is entirely optional - you can copy and paste the GPT's system instructions below into your own GPT.If you are creating your own modified version of this GPT, here are the supporting files you will need to upload into your GPT’s knowledge.Image Prompting Guide Inputs I may provide: Brand kit, audience, asset type, dimensions, visual examples, usage channel, do/don't references. Operating instructions: - First, restate the objective in one clear sentence. - If critical information is missing, ask up to 5 focused questions. If there is enough information to proceed, make practical assumptions and label them. - Use a Standard response style. - Be specific to the business, audience, channel, and constraints provided. - Avoid generic AI advice. Give concrete recommendations, examples, templates, copy, or steps I can use. - When current facts, competitors, laws, prices, policies, or market claims matter, use current research and cite sources. - Do not expose hidden chain-of-thought. Provide a concise rationale or decision summary instead. - End with a short QA checklist that helps me verify the output. Required output: Creative brief with positioning, art direction, layout guidance, asset specs, QA checklist. Caution: Avoid generic output; require concrete examples, assumptions, and next steps.
QA Follow-Up Checklist
After the AI returns its output, verify against:
- Output is specific to the provided business/context.
- Assumptions are clearly labeled.
- No unsupported claims without source checks.
- Next actions are clear and usable.
Follow-Up Prompt
Now turn the result for 'AI Image Prompt Writer' into a client-ready version: tighten wording, remove fluff, add missing assumptions, and provide the next 3 actions.
Avoid / Cautions
Avoid generic output; require concrete examples, assumptions, and next steps.
How Different Verticals Use This Workflow
Restaurant & Hospitality
A new ramen shop in Brooklyn needs hero imagery for their site launch but hasn't booked a photographer. They feed in their brand kit (muted indigo, off-white), the visual reference of Ippudo's NYC site, and the use case of homepage hero at 16:9. Output: 5 prompts for steam-rising bowls in editorial low-light, shot at 50mm. Negative prompts kill 'plastic noodles' and 'fake steam'. They use them in Midjourney and ship 3 hero options in an afternoon.
Retail & E-commerce
A men's grooming brand launching a new beard oil for the holiday window needs lifestyle scenes for Meta ads. They feed in brand colors (forest green, brass), the audience (30-45 year old men, not the bro market), and the spec of 1080x1350 for feed and 1080x1920 for stories. Output: 5 prompts each in both ratios, all featuring hands and bottle in workshop settings — no faces, no fake catchlights, no overlit kitchen counters.
Professional Services & B2B
A boutique law firm in Chicago needs a refresh of their service page imagery — currently using stock handshake photos. They feed in their brand kit (navy, ivory), the desired feel (gravitas without intimidation), and the use case at 4:3 for desktop sections. Output: 5 architectural prompts featuring courthouse exteriors and bookshelf details with cinematic underexposure. Negative prompts ban 'businessman in suit' entirely, which is the firm's actual constraint.
Beauty & Personal Care
A solo esthetician launching a new facial service needs Instagram imagery without paying for a brand shoot. She feeds in her brand kit (warm clay tones), reference of Heyday's content, and the spec of 4:5 feed and 9:16 reels. Output: 5 prompts featuring close-up product textures, golden hour skin, and water reflections — no faces, no over-airbrushed model shots that trigger the 'AI skin' tell.
Local & Trade Services
A home renovation contractor in Denver needs hero imagery for new service pages but doesn't have permission to use his client's actual remodels yet. He feeds in his brand colors (warm gray, mustard), references of Block's content, and aspect 16:9 for service page heroes. Output: 5 prompts for kitchen and bath details with morning light through windows, real material textures, no perfect staging. Negative prompts kill 'unrealistically clean construction site'.
Frequently Asked
What does a great output for this look like specifically?
Five image prompts, each with subject lock, lighting language a photographer would actually use ('soft key light from camera-left, 3:1 ratio'), aspect ratio matched to platform, negative prompts that kill the obvious AI tells (extra fingers, plastic skin, overlit catchlights), and one brand-color anchor. If the output reads like 'a beautiful product in stunning lighting,' it failed. The test: hand the prompt to three different generators and get visually consistent results. If the outputs diverge wildly, your prompt was too vague.
How is this different from product-images-from-uploaded-photo or nano-banana?
This prompt writes the prompt. The other two execute against an uploaded image. Use this when you have no source image and need to generate from scratch — concept art, mood boards, lifestyle scenes. Use product-images-from-uploaded-photo when you've already got the product shot and want to restage it in a luxury context. Use nano-banana when you need to composite your product with a model. They're upstream/downstream tools, not substitutes.
What's the most common failure mode here?
Skipping the aspect ratio question. People want a 'beautiful image' and end up with a 1:1 square they can't use in a story or 9:16 reel. Force the prompt to ask which platform first, then output the spec for that platform. Second failure: not specifying negative prompts. Without them, you get the generic glossy-AI look — extra reflections, impossible perspective, weird hands on any person. Negative prompts are 30% of the quality difference and most users skip them.
Should I use ChatGPT or Claude to write the prompts?
Claude for prompts going into Midjourney or Flux — its descriptive prose holds composition better. ChatGPT for prompts going into Canva AI or Adobe Firefly because they're tuned for shorter structured inputs that GPT writes natively. For Nano Banana specifically, GPT writes cleaner JSON. Don't agonize over this — the bigger lift is the inputs you give either model. A bad prompt in any model still produces AI sludge.