Ready-to-Use AI Image Prompts 2026: 70+ Examples to Copy

Last updated: September 2026

The center of gravity in AI image prompts has shifted. In 2026 the fastest-growing demand is not writing new descriptions at all — it is editing existing photos, and dedicated Nano Banana prompt libraries now run past 600 entries, almost all of them editing instructions. OpenAI shipped ChatGPT Images 2.5 in September 2026 with richer built-in editing. xAI launched Grok Imagine 2.0 in August. Alibaba's Qwen Image 2.1 arrived in September as a free open-weight model with RGBA transparency and support for up to ten reference images.

Most prompt lists still target the early-2026 toolset and stop at Midjourney. This page was rebuilt for the current stack and for the workflows professionals actually run: generation for products, portraits and ads, plus a full editing layer and an image-to-prompt workflow for reproducing any reference look.

You get the five-element formula, a model matrix, and 70+ ready-to-use prompts organized by job. Every prompt is written to be pasted directly, then customized with your own variables — no translation, no filler.

AI Overview

A ready-to-use AI image prompt is a reusable instruction that fixes five variables — subject, style, lighting, background and aspect ratio — so the model makes fewer guesses and returns usable results faster. Keep one template, change one variable at a time, and pair it with editing prompts when you start from an existing photo instead of a blank canvas.

The five elements of a strong image prompt

A weak prompt outsources dozens of decisions to the model. A strong prompt makes those decisions once and reuses them.

When someone types "a nice photo of a coffee bag," the model must guess the background, the camera angle, the lighting, the mood and the format. A professional prompt removes that ambiguity:

Matte black coffee bag, editorial product photography, soft directional light from the left, warm concrete surface, centered composition with copy space above, 4:5 aspect ratio

The five slots that matter:

1. Subject — who or what, with two or three concrete attributes. Not "a businessman" but "a founder in his forties, gray suit, standing in a glass-walled office."
2. Style — one visual genre: photorealistic, editorial photography, cinematic still, digital illustration, minimalist product photography.
3. Lighting — soft diffused light, dramatic side lighting, golden hour, studio strobe, rim lighting. Lighting decides mood and realism more than any other slot.
4. Background or environment — clean white background, dark velvet backdrop, marble tabletop, modern office, rain-soaked street.
5. Aspect ratio — 1:1 for catalogs, 4:5 for feed posts, 9:16 for stories and reels, 16:9 for banners and landing pages, 3:4 for portraits.

The working formula:

[Subject with 2–3 attributes] + [one style] + [lighting] + [background] + [composition] + [finish] + [aspect ratio]

One style per prompt. If you want three styles, write three prompts.

The 2026 image model matrix

Prompt syntax barely differs between tools in 2026 — behavior does. Editing strength, text rendering and reference handling now separate the models more than wording does.

Practical takeaways from running the same briefs across these tools:

• Editing is where Nano Banana and ChatGPT Images 2.5 earn their keep. If your job starts with an existing photo, start there.
• Readable in-image text is still the weakest link. Ideogram handles it best; Midjourney needs careful prompting. For campaigns, generate with "negative space for headline" and add the real text in your design tool.
• Reference-image support decides consistency. Qwen Image 2.1's ten-reference ceiling makes it the budget option for keeping one product or character stable across a batch.

The master template, filled three ways

Copy the template, then swap only the slots that change for your use case.

Template:

[Subject], [style], [lighting], [background], [composition], [finish], [aspect ratio]

Perfume ad:

Amber glass perfume bottle, luxury editorial product photography, dramatic rim lighting with gold reflections, black velvet background, centered composition with copy space above, ultra-detailed, 4:5 aspect ratio

LinkedIn headshot:

Portrait of a marketing director in her thirties, professional corporate headshot, soft diffused studio lighting, neutral gray backdrop, sharp focus on the eyes, magazine-quality finish, 3:4 aspect ratio

Interior:

Contemporary living room with slatted wood paneling, architectural interior photography, soft ambient lighting with warm accents, wide-angle composition, highly detailed, 16:9 aspect ratio

Ready-to-use AI image prompts: products and e-commerce

1. E-commerce hero shot of {product}, floating centered on a seamless pastel gradient backdrop, soft studio lighting with a gentle drop shadow, crisp catalog-ready detail, 1:1 aspect ratio.
2. Luxury {product} on black velvet, single hard key light from the upper left, deep shadows, gold reflections, premium editorial mood, 4:5.
3. {product} flat lay on white marble, styled with eucalyptus sprigs and linen, diffused morning window light, minimalist skincare-brand aesthetic, 1:1.
4. Perfume bottle macro, glass refractions, dramatic rim lighting against charcoal, fine water droplets on the glass, ultra-detailed, 4:5.
5. Tech gadget render, levitating at a three-quarter angle, dark blue gradient with subtle light streaks, futuristic premium finish, 1:1.
6. Handmade ceramic {product} on a rustic oak table, warm side light, visible glaze texture, artisan brand feel, shallow depth of field, 4:5.
7. Subscription box shot, open lid with contents arranged in a tidy grid, bright even lighting, clean studio background, direct-response look, 4:5.
8. {product} held by a smiling customer in a bright kitchen, natural light, authentic user-generated feel, 9:16.
9. Jewelry macro of {item} on dark silk, single spotlight, sparkling facet detail, auction-house aesthetic, 4:5.
10. Frozen dessert hero with visible frost, moody dark background, hard side light, indulgent food-ad styling, 4:5.
11. Sustainable {product} line-up, three variants in a row, recycled-paper texture, soft daylight, honest eco-brand tone, 16:9.
12. Split composition showing {product} boxed and in use, clean white studio, instructional e-commerce look, 16:9.

For a deeper library focused on selling angles, our AI product photography prompts expands this category with platform-specific variants.

Professional headshot and portrait prompts

13. Professional LinkedIn headshot of {person}, navy blazer over a white shirt, soft diffused key light with a subtle rim light, neutral gray backdrop, sharp focus on the eyes, 3:4.
14. Corporate portrait of a {industry} executive in a glass-walled office, natural window light, confident relaxed expression, editorial business photography, 3:4.
15. Startup founder portrait in a bright coworking space, candid mid-laugh moment, warm afternoon light, documentary style, 4:5.
16. Creative professional headshot, studio backdrop in a brand color, single-source dramatic lighting, calm confident expression, magazine profile style, 3:4.
17. Expert headshot of {person} in front of a blurred bookshelf, soft even lighting, approachable and authoritative, 4:5.
18. Healthcare professional headshot, white coat, modern clinic softly out of focus, clean trustworthy lighting, 3:4.
19. Real-estate agent portrait at a property entrance, golden hour light, friendly open posture, premium local-business branding, 4:5.
20. Finance professional headshot, charcoal suit, blue-gray backdrop, crisp studio lighting, institutional tone, 3:4.
21. Personal-brand portrait at a desk with a laptop, warm home-office light, natural gesture mid-sentence, 4:5.
22. Conference speaker on stage, soft bokeh audience lights, subtle edge lighting, energetic professional mood, 3:4.
23. Remote-team profile portrait, ring-light look with a laptop screen glow, friendly direct gaze, 1:1.
24. Black-and-white portrait of a craftsman at work, strong side light, textured hands in frame, documentary fine-art style, 3:4.

Advertising and social creative prompts

25. Facebook ad creative for {product}, bold hero product on a split-color background, clear negative space on the left for a headline, high contrast, 4:5.
26. Instagram story ad, {product} centered in the lower third, vibrant gradient background, space reserved at the top for a promo code, 9:16.
27. LinkedIn single-image ad for {service}, muted corporate palette, one clear visual metaphor, generous copy space, 1:1.
28. TikTok first-frame visual for {topic}, high-contrast subject, bold arrow pointing at the key detail, curiosity-gap composition, 9:16.
29. Retargeting ad visual, extreme close-up on {product detail}, dark background, single spotlight, premium urgency mood, 4:5.
30. Carousel cover for a {topic} guide, space for "01" numbering, editorial grid layout, muted paper texture, 4:5.
31. YouTube-thumbnail-style expressive portrait of {person}, exaggerated reaction, saturated complementary colors, strong outline, 16:9.
32. Landing page hero for {service}, subject on the right third, open headline space on the left, soft studio light, 16:9.
33. Seasonal sale banner background, abstract geometric shapes in brand colors, subtle gradient, no text elements, 16:9.
34. UGC-style product moment, phone-camera look with slight motion blur, hand holding {product} in a real kitchen, imperfect and authentic, 9:16.

Every ad prompt above reserves space for copy on purpose. Our advertising image prompts collection covers campaign-specific variants if you need more.

Food, hospitality and interior prompts

35. Overhead {dish} spread on a dark stone table, ceramic bowls, warm candlelight, rich saturated color, editorial food photography, 4:5.
36. Pour shot of {drink} into a glass with ice, backlight catching the liquid, dark bar background, high-speed capture look, 9:16.
37. Bakery hero of {item} on a floured wooden board, warm golden light, faint steam, artisan rustic mood, 4:5.
38. Meal-prep flat lay, five containers in a neat grid, bright natural light, fresh vibrant produce, 1:1.
39. Restaurant menu hero, {dish} plated center-frame on matte ceramic, single dramatic spotlight, dark elegant background, 16:9.
40. Street-food scene at night, hands preparing {item}, neon signage bokeh, steam and motion, documentary style, 4:5.
41. Espresso extraction close-up, crema detail, macro lens feel, warm side light, specialty-café branding, 4:5.
42. Modern living room interior, warm neutral palette, floor-to-ceiling windows with sheer curtains, soft afternoon light, architectural-magazine quality, 16:9.
43. Minimalist home office, pale wood desk, single plant, clean white walls, bright even daylight, 16:9.
44. Boutique hotel suite, layered textiles, brass accents, moody bedside lamps, hospitality editorial photography, 16:9.
45. Luxury villa exterior at dusk, warm interior lights glowing, pool reflection in the foreground, twilight sky, 16:9.
46. Retail store interior, warm track lighting over wooden shelving, inviting boutique atmosphere, wide-angle commercial photography, 16:9.

Editorial and cinematic prompts

47. Cinematic still, lone figure on a rain-slicked street at night, neon reflections, anamorphic flare, moody thriller grade, 21:9.
48. Editorial fashion shot, sculptural monochrome outfit against a concrete wall, hard midday sun, bold shadow shapes, 3:4.
49. Golden-hour portrait in an open field, backlight halo, soft haze, warm film emulation, nostalgic campaign mood, 3:4.
50. Double-exposure portrait merging a silhouette with a forest canopy, fine-art minimalism, muted tones, 3:4.
51. Architectural abstract, extreme close-up of facade geometry, black and white, fine-art print quality, 1:1.
52. Desert wide shot at dawn, lone vehicle track across dunes, cold blue-to-pink gradient sky, epic scale, 21:9.
53. Product on a plinth with a painted canvas backdrop, theatrical spotlight, gallery campaign aesthetic, 4:5.
54. Street-style candid, subject crossing a busy intersection in a bold coat, motion around a sharp subject, overcast light, 4:5.
55. Workspace still-life, notebook, coffee and glasses by a window, low warm lamp light, quiet cinematic mood, 3:4.
56. Surreal concept image, {object} floating above a calm sea, hyperreal render, pastel sky, dreamlike minimalism, 16:9.

If stylized, high-end visuals are your whole brief, our luxury AI image prompts for Midjourney pack goes deeper on that aesthetic than this general library.

AI image editing prompts for Nano Banana and ChatGPT Images 2.5

This is the layer most prompt lists still skip, and it is where 2026 tooling is strongest. Editing prompts work differently: you start from an uploaded photo, and the instruction controls one change at a time.

57. Replace the background of this product photo with a seamless light-gray studio backdrop. Keep the product, its shadows and reflections exactly as they are.
58. Change the outfit in this portrait to a tailored navy blazer over a white shirt. Preserve the face, hair, pose and lighting.
59. Merge these two photos into a double-exposure portrait: the silhouette from the first image, the forest texture from the second, blended at 50 percent.
60. Extend this image to a 16:9 canvas by continuing the background naturally. Do not duplicate any objects.
61. Remove the {object} from this photo and reconstruct the background behind it realistically.
62. Relight this product shot as if lit by soft window light from the left. Keep the composition unchanged.
63. Fix the distorted text on the sign so it reads exactly: "{your text}".
64. Add clean negative space in the upper third of this image for a headline, gently blurring the background only.
65. Place this packaged product realistically on the marble counter in the second image, matching perspective and lighting.
66. Color-grade this photo to a warm cinematic look: lifted blacks, soft highlights, slight teal in the shadows.
67. Clean-up pass on this old family photo: remove scratches and noise, restore facial detail, keep it natural rather than smoothed.
68. Turn this flat product photo into a 4:5 lifestyle ad: soft gradient backdrop, realistic drop shadow, and copy space at the top.

Two habits make editing prompts reliable: one instruction per generation, and explicit "keep" clauses for everything that must not move.

Image-to-prompt: reverse-engineer any reference photo

When a client sends a reference and says "like this," do not guess. Turn the image into a prompt:

1. Upload the reference to any multimodal chat model (ChatGPT, Gemini, Grok all qualify).
2. Ask it to describe the image as a generation prompt covering subject, style, lighting direction and quality, background, camera angle, color palette and aspect ratio.
3. Save the output as a reusable template with the subject swapped out.
4. Batch-check consistency: models with reference-image support, such as Qwen Image 2.1 with up to ten references, can hold the look across a whole set.

A scaffold prompt that works across tools:

Describe this image as an image-generation prompt. Specify: subject and its key attributes, photographic or art style, lighting direction and quality, background, camera angle and lens feel, color palette, and aspect ratio. Output only the prompt.

Keeping one face or brand consistent

Consistency breaks when everything changes between generations. Fix the identity slots, then vary only the scene:

{Person}, early thirties, oval face, dark brown eyes, shoulder-length black hair, navy blazer, warm confident expression — repeat verbatim in every prompt.

Derived shots from the same identity block: close-up headshot, seated at a desk, presenting to a team, looking to camera, on a video call. Change one element per version and the series stays recognizable.

If you are rebuilding this workflow for a team, ArWriter gives you one workspace where prompt templates, brand notes and drafts live together instead of scattered across chat threads.

What testing across the 2026 models taught us

Honest findings from running these briefs across the current toolset:

• Quality words are not quality control. "8K, ultra-detailed, masterpiece" nudges the finish but guarantees nothing. Lighting and composition language moves results far more.
• Changing five things at once hides the lesson. If a new version is better, you will not know why. One variable per iteration.
• In-image text is still a gamble. Outside of Ideogram and the strongest 2.5-class models, plan for "copy space" plus real typography in your design tool.
• Contradictory styles cancel out. "Luxury editorial cartoon photorealism" returns mush. Pick one genre per prompt.
• Aspect ratio decided late is wasted work. A perfect square that must become a 9:16 story means regenerating. Decide the destination first.
• Check commercial terms before client delivery. Each platform licenses outputs differently, and open-weight models you self-host put that responsibility on you.

Build a prompt library, not a prompt stash

The difference between solo and team use is organization. Classify prompts by goal (sell, explain, build trust), channel (feed, story, landing page), category (product, portrait, interior) and style. Then name them so anyone can find and extend them:

• ad-skincare-instagram-4x5-softlight-v1
• headshot-founder-linkedin-3x4-neutral-v2
• hero-realestate-landing-16x9-sunset-v3

A named, versioned library turns prompt writing from a repeated creative act into a maintainable asset. Teams that adopt this stop rewriting the same brief every sprint.

FAQ

What makes a good AI image prompt?

Specificity across five slots: a defined subject, one visual style, stated lighting, a background, and an aspect ratio matching where the image will be published. A good prompt removes decisions the model would otherwise make randomly, and stays short enough that every word controls the output.

How do I write prompts for Nano Banana?

Talk to it like an editor, not a director. Start from an uploaded photo and instruct one change at a time — "replace the background with X, keep the product and shadows unchanged." The upload fixes the subject, so your words control placement, style and corrections rather than describing everything from scratch.

What are the best AI image prompts for product photos?

Anchor the product, then fix the scene: "E-commerce hero shot of {product}, seamless backdrop, soft studio lighting, centered composition, 1:1." For ads, add "negative space for a headline" so design can overlay copy. Square for catalogs, 4:5 for social feeds, 16:9 for landing pages.

Can I use Arabic prompts in Midjourney?

Results are inconsistent. Many models parse Arabic descriptions, but Midjourney generally responds more reliably to English prompt strings, and non-Latin text rendered inside the image is unreliable in most tools. If you plan in Arabic, keep sentences simple and pair them with an English prompt string for generation.

How do I turn an image into a prompt?

Upload the reference to any multimodal chat model and ask it to output subject, style, lighting, background, camera angle and aspect ratio as a generation prompt. Save that as a template, swap the subject, and the look becomes reproducible across a whole campaign — the workflow section above shows the exact scaffold.

What is the best AI image generator in 2026?

There is no single winner — the job decides. Editing-heavy work points to Nano Banana or ChatGPT Images 2.5; stylized aesthetics to Midjourney V8; realism to Flux 2; transparent assets and reference-consistent batches to Qwen Image 2.1; readable in-image text to Ideogram. Choose by output use, not brand.

How long should an image prompt be?

Long enough to fix the five core variables, short enough that nothing contradicts — roughly 20 to 60 words in practice. That covers a sentence of subject detail, a style, a lighting direction, a background, a composition note and the ratio. Padding adds little once those slots are filled.

Do "8K" and "ultra-detailed" actually improve results?

They nudge the finish but guarantee nothing. Model choice, lighting language and composition drive quality far more than resolution words. Keep one quality tag if you like its effect, and spend the remaining prompt space on specifics: materials, lens feel, color palette.

Start from a template, not a blank box

The prompt is not the art — the system behind it is. A strong template cuts trial-and-error, keeps a campaign visually coherent, and makes each new image cheaper to produce than the last. Copy one prompt from this page, run it, change a single variable, and keep what works. Ten iterations in, you will have a template of your own — and that template, not any single image, is the asset worth maintaining.

Sources

• https://openai.com/news/
• https://x.ai/news
• https://qwen.ai
• https://help.openai.com/en/articles/10032626-prompt-engineering-best-practices-for-chatgpt
• https://docs.midjourney.com/docs/aspect-ratios

Try ArWriter today

Turn this prompt library into a repeatable production workflow — templates, brand notes and drafts in one workspace.

Start free →