Good image prompts are easier to write when you begin from a repeatable pattern and adapt it to the job. This guide organizes common visual requests into reusable structures, then shows which details should change for portraits, products, posters, illustrations, and edits.
The useful principle
Describe relationships and priorities, not only a bag of keywords.
Say where the subject appears, what the viewer should notice first, how elements relate, and which details must remain exact. Use editing language for revisions instead of rewriting everything from zero.
What you can use these prompts for
Start with the outcome you need, then add the information that changes the answer. These are common use cases—not rigid formulas—and each one should be adapted to your real material.
Portraits
Define person, expression, pose, lens feeling, light, background, and identity constraints.
Product shots
Protect geometry and branding while controlling surface, camera angle, reflections, and space for copy.
Posters
Design the text hierarchy and composition together so neither becomes an afterthought.
Illustrations
Describe narrative moment, medium, shape language, palette, and intended audience.
Social carousels
Keep a stable visual system across slides while changing scene and message.
UI concepts
State device, layout, components, content density, state, and brand direction.
Diagrams
Prioritize factual relationships, simple labels, legibility, and a restrained visual style.
Image edits
Separate the requested change from everything that must remain untouched.
A reliable way to build the prompt
1. Choose the visual job
Decide whether the image must explain, persuade, identify, decorate, or tell a story. The job determines what deserves visual priority.
2. Select one strong hierarchy
Name the focal point, supporting elements, and background. Too many competing subjects make a visually busy but unclear result.
3. Translate style into visible properties
Instead of only saying ‘cinematic’ or ‘premium,’ describe contrast, light, material, spacing, color restraint, lens feeling, and finish.
4. Keep a revision log
Change one variable at a time and record what improved. Reusing the successful structure is more reliable than accumulating random adjectives.
Prompt templates you can copy and adapt
Replace every bracketed placeholder with real information. Keeping placeholders visible is safer than letting an AI invent missing facts.
Create a vertical editorial portrait of [subject]. Expression: [emotion]. Wardrobe: [details]. Frame from [crop] using a natural eye-level perspective. Background: [environment], softly separated from the subject. Lighting: large soft key from camera left with a subtle violet edge light. Color grade: restrained, realistic skin tones, deep navy shadows. Leave negative space on the right for article text. No embedded words.
Create a cohesive five-image Instagram carousel about [topic]. Use the same dark navy world, violet light trail, typography hierarchy, and recurring [character/object] on every slide. Each slide must have a different clear scene and one short headline. Keep characters and brand colors consistent, preserve generous margins, and design for 4:5 mobile viewing. Slide messages: [list exact messages].
Edit the supplied image. Change only [specific element] to [desired change]. Preserve the character identity, facial expression, pose, camera angle, composition, lighting direction, background geometry, logo, and all existing text exactly. Match the original rendering style and image quality. Do not add new objects.
Why specificity changes the result
Generate a beautiful social media image.
Create a 4:5 Instagram cover about five prompt mistakes. Use a deep navy background, one luminous violet path leading toward a confused paper character, and a bold white headline in the upper third: “5 mistakes weakening your AI answers.” Keep the composition readable on a phone, use one focal point, and add no other text.
Quick quality checklist
✓The image’s purpose and aspect ratio are defined.
✓Subject, action, environment, and composition are explicit.
✓Lighting and visual language support the message.
✓Exact text and placement are specified only when needed.
✓Protected elements and unwanted elements are clearly named.
Use AI as an assistant, not an authority
Generated images may contain incorrect text, anatomy, product details, trademarks, cultural stereotypes, or visual similarities you did not intend. Inspect every asset before publication, verify commercial rights and platform rules, and obtain consent for real people’s likenesses or sensitive personal material.
The takeaway
Start from a reusable visual pattern, then make the subject, hierarchy, light, text, and protected details specific to the image’s real job.