How we tested this
Reviewed against Leonardo AI documentation in May 2026
We refreshed this guide by checking Leonardo's current documentation for image generation, Image Guidance, model availability, prompt parameters, video generation, and commercial-use language on the pricing page.
The prompts below are written for repeatable visual production rather than one-off novelty art. They include style control, negative prompts, reference-image strategy, product safety checks, and game-asset constraints.
How to prompt Leonardo AI
Leonardo AI responds well to prompts that name visible details in a stable order. Start with the output type and style, then describe the subject, setting, camera, lighting, and any production constraints. Add a negative prompt for the mistakes you see most often.
Prompt map
Leonardo AI workflows and starter prompts
Pick the workflow before writing the prompt. Character art, product images, game assets, and reference sets need different constraints.
| Job | Prompt focus | Starter prompt | Human check |
|---|---|---|---|
| Character concept | subject, silhouette, costume, pose, reference strength | Create a full-body character concept of [character]. Style: [style]. Include [costume details], readable silhouette, [pose], [background], and consistent lighting. Negative prompt: extra limbs, distorted hands, unreadable face, text. | Use Character Reference only for consistency, not as a face-swap guarantee. |
| Product visual | material, camera, lighting, background, text space | Create a premium product image of [product] made from [materials]. Use [camera angle], soft studio lighting, clean background, realistic reflections, and negative space for headline copy. | Check brand accuracy, logo placement, label text, and any implied product claims. |
| Game asset | asset type, view angle, style, transparency, scale | Create a [asset type] for a [game genre] game. View: [top-down/isometric/side]. Style: [style]. Use a clear silhouette, consistent light direction, simple background, and game-ready detail. | Test readability at the size the asset will appear in game. |
| Style reference set | shared style, palette, motif, reference weight | Generate 6 visuals in one consistent style for [project]. Palette: [colors]. Motif: [motif]. Keep the same lighting, lens feel, texture, and composition rules across every image. | Save the strongest output as a style reference for the next batch. |
| Ad creative | buyer, scene, benefit, composition, compliance | Create an ad background for [offer]. Audience: [buyer]. Show [benefit] through a realistic scene, leave room for headline and CTA, and avoid text, logos, medical claims, or exaggerated results. | Add final copy and proof manually in your design tool. |
Copy this first
The all-purpose Leonardo AI prompt
Leonardo AI character prompts
Use character prompts for concept art, avatars, comic characters, RPG portraits, and game prototypes. Consistency improves when you keep silhouette, costume, palette, and camera language stable.
Prompt 1
Full-body concept art of a cyberpunk courier, reflective jacket, compact backpack, rain-soaked alley, neon signage glow, confident stance, cinematic close detail, negative prompt: extra fingers, blurry face, text.
Prompt 2
Fantasy healer character, linen robes, brass medical satchel, glowing botanical staff, mossy forest path, soft dawn light, painterly illustration, clear silhouette.
Prompt 3
Stylized 3D mascot for a productivity app, friendly robot assistant, rounded shapes, teal and white palette, simple studio background, expressive eyes, brand-safe design.
Prompt 4
Anime-style academy rival, tailored black uniform, silver hair, sharp expression, library interior, dramatic rim light, half-body portrait, clean linework.
Prompt 5
RPG merchant NPC, layered fabrics, many small pouches, desert marketplace, warm sunlight, grounded realistic fantasy style, three-quarter view.
Prompt 6
Comic-book superhero for a climate campaign, wind-powered suit, green and blue accents, heroic pose, city skyline, bold ink lines, high contrast.
Prompt 7
Character turnaround prompt for [character], front view, side view, back view, same outfit, same lighting, neutral background, production sheet style.
Prompt 8
Portrait of [character] with a consistent face shape, hairstyle, costume colors, and mood. Use Character Reference if available and keep reference strength at a flexible level first.
Product and brand image prompts
For product visuals, Leonardo works best when you describe material, reflections, lens, surface, and ad-space. Do not rely on AI to create accurate labels or logos.
Prompt 1
Premium studio photograph of a matte black smartwatch on dark stone, softbox reflection, shallow depth of field, subtle blue rim light, centered composition.
Prompt 2
Eco skincare bottle on white marble, frosted glass, eucalyptus leaves, soft daylight, clean shadows, luxury wellness aesthetic, no readable text.
Prompt 3
Wireless earbuds floating above case, exploded-view composition, white background, subtle reflections, technology marketing style, crisp detail.
Prompt 4
Coffee subscription hero image, kraft paper bag, ceramic mug, roasted beans, morning window light, warm neutral palette, negative space on left.
Prompt 5
SaaS dashboard hero background, laptop at clean desk, blurred analytics charts on screen, soft studio light, professional B2B mood, no fake UI text.
Prompt 6
Packaging concept for [product], simple geometric label area, bright shelf-ready colors, realistic cardboard texture, front-facing mockup, no brand logo.
Prompt 7
Lifestyle ad scene for [offer], [audience] using [product] in [setting], natural expression, realistic lighting, space for headline, commercial photography.
Prompt 8
Create 5 visual directions for [brand]. Each should include product placement, surface, lighting, color palette, and what emotion the image should create.
Game asset and concept art prompts
Game prompts need production constraints: camera angle, asset type, readable shape, scale, and consistent lighting. Add transparency or plain backgrounds when the asset will be cut out later.
Prompt 1
Isometric fantasy house asset, blue roof, stone chimney, warm window glow, clean edges, mobile game style, plain background, consistent top-left lighting.
Prompt 2
Top-down dungeon floor tileset, cracked stone, moss variations, torchlight stains, seamless pattern, readable at small scale, dark fantasy palette.
Prompt 3
2D side-scroller platform asset, floating grass island, chunky shape language, bright colors, simple shadows, transparent-style plain background.
Prompt 4
Inventory icon for a magical healing potion, glass bottle, red liquid, cork stopper, gold trim, centered, high contrast, game UI icon style.
Prompt 5
Sci-fi weapon concept sheet, compact plasma pistol, three angle views, labels area left blank, gunmetal material, cyan energy core, clean background.
Prompt 6
Trading card frame for a fantasy game, ornate gold border, central empty art window, bottom text panel blank, high contrast, premium collectible style.
Prompt 7
Creature concept for [game biome], readable silhouette, 3 unique features, neutral pose, material callouts, natural habitat background.
Prompt 8
UI button set for [game genre], normal state, hover state, disabled state, consistent bevel, same light direction, no text inside buttons.
Style reference and Image Guidance prompts
Leonardo's Image Guidance can use style, character, content, edge, depth, and pose references depending on model support. Prompt the intended relationship between the reference and the new output.
Prompt 1
Use this image as a Style Reference only. Keep its palette, texture, lighting, and lens feel, but create a new subject: [subject]. Do not copy the original composition exactly.
Prompt 2
Use this character reference for overall identity consistency. Generate [new scene]. Keep hairstyle, costume colors, and face shape similar, but allow a new pose and background.
Prompt 3
Use this content reference for layout. Create a new [asset] with the same composition and spacing, but change the style to [style].
Prompt 4
Use this pose reference for a character illustration. Keep the body pose and camera angle, but create a new outfit, setting, and lighting mood.
Prompt 5
Generate 4 variations from this image. Keep the main composition and style, but test different lighting: golden hour, studio softbox, moonlight, and overcast daylight.
Prompt 6
Create a matching set of 8 images using the same style reference. Keep color palette, texture, lens, and contrast consistent across the full batch.
Prompt 7
Use low reference strength first to preserve prompt freedom, then increase reference strength only if the style or character drifts too far.
Prompt 8
Audit this prompt and reference plan. Tell me whether I should use Style Reference, Character Reference, Content Reference, Pose Reference, or no reference.
Negative prompts and repair prompts
Negative prompts are not magic, but they help remove repeated failure modes. Keep them targeted, then use repair prompts for composition, anatomy, text, and brand issues.
Prompt 1
Negative prompt for product images: misspelled text, fake logo, warped label, extra product, duplicate bottle, dirty surface, low resolution, unrealistic reflection.
Prompt 2
Negative prompt for character art: extra limbs, extra fingers, fused hands, distorted eyes, asymmetrical face, unreadable costume, blurry details, watermark, text.
Prompt 3
Negative prompt for game assets: cropped asset, inconsistent perspective, cluttered background, unreadable silhouette, muddy colors, text, UI labels, watermark.
Prompt 4
Repair this prompt so the result has clearer hierarchy, stronger lighting, fewer objects, and more negative space for design copy: [prompt].
Prompt 5
Make this image prompt more commercially safe. Remove celebrity likeness, trademark references, copyrighted characters, and any claim the image cannot prove.
Prompt 6
Create a second-pass prompt that keeps the subject but changes only the camera angle, lighting, and background mood.
Prompt 7
The output looks generic. Add 8 concrete details about material, setting, lens, texture, and composition without making the prompt longer than 5 sentences.
Prompt 8
The character is inconsistent across images. Rewrite the prompt as a reusable character bible with face, hair, outfit, palette, silhouette, and pose constraints.
What to review before publishing Leonardo AI images
Leonardo AI prompt FAQs
What is the best prompt format for Leonardo AI?
A strong Leonardo AI prompt describes the style, subject, subject details, scene, composition, lighting, and constraints. For production work, add model or preset intent, reference-image instructions, and a short negative prompt. The goal is to describe what the image should visibly contain, not just what the image is for.
Should I use negative prompts in Leonardo AI?
Yes, but keep them focused. Use negative prompts for common problems such as extra fingers, warped labels, fake text, watermarks, low resolution, duplicate objects, and cluttered backgrounds. Very long negative prompts can become noisy. Start with 6 to 12 clear exclusions and refine from the output.
How do I get consistent characters in Leonardo AI?
Use a stable character description, repeat the same costume and color details, save successful outputs, and use Character Reference when available. Leonardo's docs note that Character Reference is not intended as a face-swap feature and does not guarantee a perfect replica, so use it for consistency, not identity cloning.
What is Image Guidance in Leonardo AI?
Image Guidance lets you use uploaded or generated images as references. Leonardo documents options such as Style Reference, Character Reference, Content Reference, Edge to Image, Depth to Image, Pose to Image, and other ControlNet-style inputs depending on model support. It is useful when text alone cannot control style or composition.
Which Leonardo AI models should I prompt for?
Leonardo's current model list includes Leonardo models such as Lucid Origin, Lucid Realism, and Phoenix, plus other image models such as Flux, Seedream, Nano Banana, and GPT Image options. Availability can change by account, plan, and rollout. Choose the model based on output style, then adapt the prompt to that model's strengths.
Can Leonardo AI make game assets?
Yes. Leonardo AI is especially useful for concept art, item icons, isometric assets, tile concepts, character sheets, UI frames, and visual exploration. For actual production assets, check scale, silhouette, transparency needs, perspective, and consistency before bringing the image into a game engine.
Can I use Leonardo AI images commercially?
Leonardo's pricing page says commercial-use rights depend on subscription status. Paid subscribers retain ownership and intellectual property rights for generated images, while free-tier users receive a non-exclusive, royalty-free licence for commercial use and their public assets may be accessible under service features. Check the current terms before client or resale use.
How should I prompt Leonardo AI for product images?
Describe the product material, surface, lighting, camera angle, background, reflection, and space for copy. Avoid relying on Leonardo AI to generate accurate label text or real logos. Generate the visual background first, then add real text, logo, and compliance copy in a design tool.
Does Leonardo AI support video prompts?
Yes. Leonardo's API docs include text-to-video generation with models such as Motion and third-party video models, depending on availability. Video uses more compute and tokens than images, so write tighter prompts: subject, action, camera motion, duration, style, and what should not happen.
What is the main mistake beginners make with Leonardo AI prompts?
The main mistake is writing a mood instead of a visual brief. Stunning fantasy art is too vague. A better prompt defines the subject, silhouette, setting, camera, lighting, style, constraints, and negative prompt. Leonardo gives better results when the prompt can be translated into visible choices.