A full-body prompt has a different job from a portrait: it must make the entire character readable at once. Pose, silhouette, armor, weapon, footwear, and the space around the figure all compete for attention. This guide shows how to control those elements without turning the prompt into an inventory list.
Choose full-body framing when the character's class fantasy depends on more than their face. A paladin's shield stance, a ranger's layered traveling gear, a monk's balance, or a warlock's staff and familiar all need space below the shoulders. Full-body images also work well for party lineups, campaign handouts, costume references, standees, and character-design sheets.
Use a portrait instead when expression and facial identity matter most. Use a token when the image will sit on a battle grid, and use a scene when the environment or action is the true subject. A full-body image can imply a story, but the character should remain the dominant visual element.
Start with a pose that can be understood as a silhouette. Strong pose verbs include striding, bracing, guarding, kneeling, drawing a bow, and leaning on a staff. Weak phrases such as "dynamic pose" ask the image model to invent the action and often produce theatrical but unclear anatomy.
Describe the direction of the torso, feet, and main prop. "Standing with weight on the rear leg, shield angled forward, sword lowered at the right side" gives the model a coherent arrangement. Avoid asking both hands to perform unrelated actions unless each action is essential.
Leave visible space around the head, hands, weapon, and feet. Terms such as "complete figure visible from head to toe," "uncropped boots," and "weapon fully inside frame" reduce accidental clipping. A vertical 2:3 or 3:4 composition usually gives a single figure enough room without making the character too small.
A readable full-body design needs one dominant shape. A broad shield, long cloak, oversized spellbook, curved bow, or pair of horns can carry the silhouette. Pick one hero prop and let secondary equipment support it. Listing a sword, axe, bow, shield, backpack, potion belt, and familiar at equal importance makes the model compete with itself.
Describe clothing from large layers to small details: outer silhouette first, then armor or robe, then belt-level accessories, then footwear. Color should follow the same hierarchy. A useful palette instruction is "deep green cloak over brown leather, with small brass accents" rather than a separate color for every object.
elf ranger full body, green clothes, bow, boots, forest, dynamic pose, detailed fantasy art
complete full-body illustration of a wood elf ranger striding forward with weight on the front foot, longbow held low in the left hand, moss-green travel cloak forming a triangular silhouette over fitted brown leather armor, quiver visible behind the right shoulder, weathered knee-high boots fully inside frame, copper hair braided away from the face, simple forest-path backdrop, soft overcast light, vertical 2:3 composition, readable fantasy character design
The revised prompt replaces the vague request for energy with a specific movement. It organizes the cloak, armor, bow, quiver, and boots by visual importance while explicitly protecting the full figure from cropping.
dwarf paladin with armor, shield, hammer, holy symbol, heroic, full body
head-to-toe character illustration of a stocky dwarf paladin in a grounded guard stance, round shield angled forward as the dominant shape, short warhammer resting against the right shoulder, layered steel plate over a dark blue gambeson, small silver sun emblem fixed at the chest, heavy square-toed boots planted apart and fully visible, braided gray beard, restrained heroic expression, plain stone-hall background, warm side light with clear fill, vertical 3:4 framing
This version gives the shield priority and places the hammer in a quieter position. The stance, boot placement, and framing instruction make the result useful as a character reference rather than an ambiguous waist-up portrait.
Natural-language pose descriptions and a vertical aspect ratio are a practical starting point. Use --ar 2:3 or --ar 3:4, then vary the result if hands or equipment overlap. --style raw can help when decorative styling overwhelms the requested costume, but results vary by current model version.
Write the prompt as a clear layout specification: one character, complete head-to-toe framing, named hand positions, and a simple background. If an item keeps being cropped, repeat its spatial requirement in plain language rather than adding more style adjectives.
Put structural tags early, such as full body, standing, head to toe, solo, vertical composition, followed by the character and costume. A pose reference, OpenPose, or another control workflow can improve repeatability when exact limb placement matters. Negative prompts can reduce cropping and duplication, but they cannot replace a coherent positive pose description.
Keep every bracket concrete, but do not force every possible detail into one prompt. Models differ in how they weight prompt order; placing framing, pose, and the hero prop early simply makes the intended hierarchy easier to understand and easier for you to revise.