4,545 templates · updated Oct 2026
Seedream 4.5 is ByteDance's image model for posters, product shots and illustrations with short on-image text. On GPT Image Hub it renders at 2K or 4K for the same 33 credits, with up to 10 reference images.
Start with the writing guide below, then pick one of the 4,545 library templates set up for Seedream 4.5 and the rest of the Seedream family (4.0, 5.0, 5.0 Pro and 5.0 Lite). Each template opens in the generator with its recommended Seedream version selected.
Collectible art toy character design sheet, front, side, and back views. Fluffy, cute character with a round face, wearing [a yellow safety helmet] and [blue round glasses]. Soft tactile plushie figure texture, cute 3D vinyl toy style. Pastel color scheme of [black, cream, and light yellow]. Clean solid beige background, soft studio lighting.
Tang-style fantasy water-element xianxia character poster, [9:16 vertical] aspect ratio. A [23-year-old] oriental female, Tide Sound Pavilion Envoy, with a slender, tall figure, natural full bust, defined waistline, and an upright, graceful posture. Her temperament is gentle, quiet, high-end, and refined. Long oval elegant face with soft features, gentle willow-leaf eyebrows, slender eyes, natural nose bridge, soft lips, and a thoughtful expression. Black hair half-tied in a low style with side volume, adorned with a green and white jade hairpin and short pearl earrings. Wearing an elegant, semi-sheer [smoky lotus] single-layer crinkle light gauze top over a high-waisted long skirt secured with a narrow silver waistband; the sheer fabric subtly reflects the lake light, showing a natural silhouette while dense wave patterns preserve modesty. She leans sideways against the wooden railing of a lake pavilion, one palm naturally resting on the wood, the other relaxed by her side. Three faint concentric water ripples spread along the lake surface around her, casting subtle spiritual energy. Setting includes the lake pavilion railing, expansive water surface, distant misty mountain rain clouds, and lakeside reeds, illuminated by soft rainy diffuse light.
A textured oil painting in [Renaissance style], depicting a female model filling the frame. Her skin is pale with [a rosy flush and subtle freckles], and her lips are a crimson hue. The background is a textured, greenish-gray stone architectural panel featuring a faint arch motif. Soft, directional lighting emphasizes the rich textures of fabric and skin. Visible canvas texture throughout.
Detailed, text-free animation production character sheet in a hard-edged retro-futurist 1980s/1990s cel-shaded anime style, arranged on a clean flat dark charcoal background in a 16:9 widescreen layout. Characters: - [KES VOSS]: A slim, short-haired young adult with [rusty orange] hair, sharp anime eyes, guarded expression, wearing a worn orange utilitarian pressure suit with black collar, harness straps, rugged boots, patched knees, a mechanical reinforced forearm brace with exposed components and a strapped pressure dial, and an analog chest gauge connected with hoses. - [DARIUS]: A broad, stocky elderly man with dark skin, short white curly hair, white mustache and stubble, heavy brow, weathered face, wearing a bulky blue-gray space survival suit, heavy gloves, suspenders, scuffs, and a torn sleeve patch. Layout and Elements: - Left side ([KES VOSS]): Full-body front turnaround, full-body back turnaround, 3/4 bust portrait, small tense profile looking down, wrist gauge closeup, and boot sole detail. - Right side ([DARIUS]): Full-body front turnaround, full-body back turnaround, 3 head expression studies (including goggle band), collar restraint ring accessory study, and heavy glove detail. - Lower edge: A cinematic horizontal strip of 4 rectangular action panels showing zero-gravity survival scenes—young character welding with bright sparks outside, cockpit/helmet interior closeup, older man repairing machinery amid debris and floating cables, and an exterior shot beside a rugged portable field terminal with a cracked green CRT screen and chunky analog buttons. Style & Atmosphere: Gritty cel-shaded concept art, clean ink linework, muted industrial palette, scuffed metal and fabric, analog space-station grime, dramatic functional lighting. Constraints: Strictly no readable text, no titles, no UI labels, no logos, no watermarks, no chibi proportions.
A side-by-side composition featuring the original person from the reference photo kept fully photorealistic and untouched, standing alongside a [3D caricature companion] version of themselves. The caricature shares the exact identity, outfit, hairstyle, accessories, and facial resemblance, rendered in a [Pixar/Disney-inspired premium 3D styling] with large expressive eyes, a slightly oversized head, smooth cinematic skin, and fashionable proportions standing at [shoulder to ear level] relative to the real person. The real human and the animated caricature interact naturally, [leaning in playfully, smiling, and looking at each other]. Preserve the original background scene, perspective, and lighting from the reference photo, seamlessly integrating the 3D character with realistic cast shadows and cinematic editorial depth.
NewA cinematic portrait of a character with clear facial features and bright, fair skin, wearing a red mohair sweater, white mink-fur hat, and white scarf. Standing on a snowy Christmas town street at night with hanging lights and red bows, snow gently falling, distant snow-covered houses softly blurred. Warm, soft lighting adds dimensionality, snow on the sweater and scarf, smiling warmly, face and scarf details sharp and clear.
NewTransform the entire scene, including the background and surroundings, into a high-budget 3D Pixar animation movie still. The background objects and environment must be rendered as stylized 3D CGI assets with soft, rounded edges and vibrant cartoon colors, replacing any photorealistic textures. The character features smooth skin, exaggerated large eyes, and cute 3D proportions. Render everything using Unreal Engine 5 with bright cinematic global illumination and ray-tracing to ensure the character and background blend perfectly into a cohesive 3D animated world.
NewTransform into a bubblegum cartoon style, retaining the character's facial features, candy color palette, thick outlines, soft and cute shapes, playful facial expression. If there is inappropriate (sensitive) content in the background, it can be modified or removed. Redrawn composition: maintain the same aspect ratio as the original image.
Convert to Irasutoya cartoon style, preserving character facial features, thick outlines, pastel flat color blocks, minimal shadows, cute expressions. If there is any inappropriate (sensitive) content in the background, it can be modified or removed. Redrawn composition: maintain the same aspect ratio as the original image.
