Z-Image is Tongyi-MAI’s image generation model, also known as 造相. Turn a written idea into a still image: explore character portraits, illustrations, graphic posters and scene compositions, then choose the aspect ratio for your layout.
Compare three text-to-image directions: a silver-haired character portrait, a red-blue-yellow poster and a dance scene lit by red lights and smoke. Read each full original prompt alongside its still-image result.
Stylized Character Portrait
The prompt combines silver-white hair, a white shirt, red thread and a gray wall. The half-body portrait shows how these color and styling cues work together.
Male, with exquisite and three-dimensional facial features, detailed facial details, deep eyebrows and eyes, dark gray pupils, thin and delicate lips, finely drawn style, silver-white medium-length hair with distinct strands, white chiffon shirt with a slightly open collar, red fine thread wound around hands and fingertips, slightly raising the head, prominent Adam's apple, pale skin, ultra-clear quality, red letters tattoo on the collarbone, realistic thick application, soft and slightly hazy, dark black, looking at the camera, sickly appearance, oil painting touch, grayish-blurred wall background, red fine thread wound around the waist, wide shoulders and narrow waist, thin skin, exposed navel outfit, distinct abdominal muscle lines, hands pointing at the abdominal muscles, close-up half-body shot, fingers with tattoos, dramatic lighting.
The prompt specifies yellow clothing, a red beret, a blue-red background and “GRACE” lettering. The result also renders some field names and instructions as text; proofread and correct the design before publication.
"platform": "Imi Dream",
"purpose": "VLOG cover poster",
"style": "retro cartoon",
"color_tone": "strong colors",
"background": "a large area of bright red, with geometric shapes of blue and red, as well as multiple patterns of cookie-shaped crackers with curves" "central_figure": {
"Description": "A fashionable woman dressed in bright yellow clothing" "accessories": [
"Red beret",
"Golden thick-framed sunglasses",
"Red bead necklace",
"Metal necklace",
"Red gloves" ],
"expression": "Exaggeratedly excited expression, with wide-open mouth and laughing loudly",
"gesture": "Made the rock gesture" },
"text_elements": {
"top": "MyFoodChannel (bold text with a yellow background and a border)" "bottom": "GRACE"
},
"additional_elements": [
"Vintage cartoon-style cookie packaging illustration" ],
"hashtag_or_label": "TRIPLECHIC",
"studio": "STUDIO",
"tagline": "Quickly create a POP/VLOG poster spell with Dream Instant" }
An industrial basement, red light, smoke and black dancewear establish the scene. The output is one still composition for exploring a performance setting or visual mood.
In the industrial-style basement with red lights and smoke, the dancers wore black crop tops and work pants, performing a powerful hip-hop dance routine. The focus was on showcasing the strength in the waist and precise hip movements. Their expressions were stern yet confident, and sweat glistened. The camera shots were fast and varied, including wide-angle low-angle shots. Style: Dark, sophisticated, and full of tension.
Z-Image is Tongyi-MAI’s image model. Its foundation version emphasizes creative freedom, aesthetic variety and diverse compositions. Use text-led generation to explore characters, illustrations and scenes for concept discussions or layout studies.
This page offers a text-to-image workflow. Z-Image-Turbo is a separate distilled variant, and Z-Image-Edit targets image editing.
Build a Character or Scene from Text
Describe appearance, clothing, environment and light together. The portrait and basement dance examples show distinct directions for character styling and scene atmosphere.
Explore Illustration, Photography and Graphic Styles
The foundation model supports varied artistic styles. Compare the illustrated portrait, photographic scene and high-contrast poster to choose a visual direction. Small lettering still needs careful review.
WORKFLOWS
Create an Image from a Text Prompt
Start in the generator at the top of this page with a text prompt; no image upload is required. Describe the subject, setting, camera view, light and visual style for a new image.
Describe the Subject, Scene, and Style
For a character concept, describe appearance, clothing and framing. For a scene, define the foreground, subject position and background. Specify any poster wording explicitly and proofread it in the result.
PROMPT GUIDE
How to Direct Z-Image with a Prompt
These two guides retain the original English prompts from their generation records. Inspect the framing, lighting and lettering in each result before adapting the prompt to your own subject.
Frame a Character and Place the Light
The prompt specifies a cafe, cat-eared maid, black-and-white outfit, red ribbon, side-window light and syrup over pancakes. The result includes these main elements, while gaze and background details differ from parts of the request.
Close-up portrait, tsundere catgirl maid in a cozy traditional-meets-modern cafe, wearing a black maid dress with white frills and a red ribbon, black cat ears and tail twitching slightly, looking away with a slight blush while carefully squeezing heart-shaped syrup onto a pancake, soft morning light from the side window illuminating her face and hair strands, shallow depth of field, background is a softly blurred bookshelf and warm wooden counter, intimate atmosphere, expressive eyes, (detailed rendering of hair and fabric:1.2), inspired by VOFAN’s lighting and character design, digital painting.
Negative prompt: poorly drawn face, bad hands, full body, wide shot, cold colors, plain background.
Specify the Words, Then Proofread the Layout
The prompt lists a title, subtitle and publication details. The result shows “NEON PUNK LOVE” and “SATIN & STUDS”, alongside misspellings, repeated words and extra field names. Treat it as a layout draft and correct the text before use.
Write your description in the generator above, choose an aspect ratio and quantity, and review the points estimate before generating. The button below returns to the generator.
Describe the Image
Describe the subject, pose, environment, composition and light. You can copy a full case prompt from this page and replace its subject or setting.
Check the Available Settings and Points
Choose from 1:1, 2:3, 3:2, 4:3, 3:4, 16:9, 9:16 and 21:9, then select quantity. Review the current points estimate and account access. Aspect ratio controls the shape; check the downloaded original for pixel dimensions and file format.
Open and Check the Result
Generate, then inspect the subject, hands, composition and every requested word. Adjust the prompt when needed and review the next result against the same brief.
Answers about this version, its inputs, and how to get started.
What is Z-Image?
Z-Image, also known as 造相, is Tongyi-MAI’s image generation model. Its foundation version focuses on creative generation, aesthetic variety and diverse compositions. This page provides text-to-image generation for still images.
What is Z-Image useful for?
Use it to explore character styling, illustration, poster layouts and scene concepts. Describe appearance, environment, color and light, then refine the result. Choose a reference-image workflow when your task depends on an existing product or person.
Is Z-Image an image editor?
This page uses text-to-image generation. To edit an uploaded product or portrait, use an image-editing workflow. Z-Image-Edit is a separately named model variant.
What input does Z-Image need?
Enter a text prompt describing the subject, setting, composition, light and style. No reference image is required here. Use a reference-based workflow when you need to retain details from an existing product or person.
Which size and aspect-ratio settings are available?
Choose 1:1, 2:3, 3:2, 4:3, 3:4, 16:9, 9:16 or 21:9. These ratios are not fixed pixel specifications. Original case files on this page include 888×1184 and 768×1360 images; previews may be smaller. Check your downloaded original for its final dimensions.
How should I write a Z-Image prompt?
Start with the subject and scene, then add camera view, light, color and style. Specify exact wording when needed and proofread the result. The two guide prompts retain their original full wording; this form has no separate negative-prompt or CFG setting.
How can I check Z-Image access and usage costs?
Allowances and model access depend on the current WeShop plan and your account. Select your settings and quantity, then review the generator’s points estimate before submitting. The charge may vary by request.
What limitations should I check?
Inspect anatomy, hands, clothing, subject placement and text. The poster examples contain misspellings, repetition and extra field names, so proofread the design before use. This page produces still images.
What can I learn from the examples on this page?
The three examples show character styling, poster color and a dance-scene composition. The two guides explain a cafe character brief and poster wording. Each image is paired with its full original prompt; one result does not guarantee identical outcomes.
How does Z-Image differ from related models?
Tongyi-MAI describes Z-Image as the foundation generator, Turbo as a distilled variant and Edit as an editing variant. Use this page for text-led image creation. Compare Nano Banana Pro for reference editing or selectable 1K, 2K and 4K output.
