grok-image

由 xAI 提供图像

用 xAI Grok Image 提交提示词并选择宽高比,生成或编辑图像后查询异步结果。

更新于:
输入提示词文本
模型1款可用
返回方式异步结果链接
价格2 quota / 张
France travel artwork inside an arched glass display with the Eiffel TowerFrance travel poster示例 01
Cat-shaped figure made from grass and red flowers in a dark forestFloral grass cat示例 02
Viking warrior charging through snow on a scarred polar bearViking polar bear charge示例 03
Small Kodama-like forest spirit floating among glowing sporesKodama forest spirit示例 04
生成示例

从提示词到成图

Vintage melancholic vibe, phone camera style, super low resolution, close-up face shot, warm Japanese-style filter. The pic shows a super fair, slim, gorgeous cool-toned, girl with messy bun hair, curly bangs,some strands covering her eyes, wearing a light white fitted long-sleeve top, covered with a blanket, leaning back while holding a persian cat. Only her eyes are visible, with her hair naturally falling around. The whole scene has this dreamy, hazy feel. It's like a photo taken with just a regular iPhone, with a bit of motion blur, kinda messy composition, just a quick snap, like a candid shot you'd take without thinking.

A close-up of a face adorned with intricate black and blue patterns. The left side of the face is predominantly yellow, with symbols and doodles, while the right side is dark, featuring mechanical elements. The eye on the left is a striking shade of yellow, contrasting sharply with the surrounding patterns. The face is partially covered by a hooded garment, cinematic_1940s cinematic_octane

An old ninja turtle is playing poker with some old decrepit humanoid rats. The ninja turtle is smoking a cigar. The room is dark and the air is cloudy with cigar smoke. The ninja turtle is running dangerously low on chips. (cinematic) Negative prompt: bad ugly jpeg artifacts

sinozick style, fiery aura, 1girl, cyberpunk, dark theme, metal mask, metal three quarters face mask, solo, cyborg, skull, looking at viewer, multi-colored hair, white eyes, bloody, large breasts, claws, "sinozick style", "g4n1m3" Negative prompt: score_5, score_4, watermark, signature, artist name, 3d, easynegative, neutral, angry, mean, intense, ugly, old, , striped face, black stripes,

A solitary astronaut standing at the edge of an enormous extraterrestrial valley, facing a colossal ancient artificial structure partially buried beneath alien terrain. The structure extends several kilometers into the distance, its scale almost incomprehensible, constructed from unknown dark metallic material covered with subtle geometric erosion patterns. A gigantic ring-shaped planet dominates the upper atmosphere while distant moons remain barely visible through atmospheric haze. The astronaut occupies less than three percent of the frame, deliberately emphasizing overwhelming scale. Camera positioned behind and slightly below the astronaut, 21mm large-format cinema lens, monumental vertical composition. Cold distant sunlight passes through suspended extraterrestrial dust, producing enormous volumetric shadows from the alien structure. Realistic spacesuit materials, fine dust accumulation around boots, physically plausible terrain, accurate atmospheric perspective. Hard science-fiction production design, restrained technology, no random glowing panels, no fantasy magic, no excessive visual clutter. IMAX science-fiction cinematography, monumental scale, photorealistic environmental concept art, sophisticated cinematic color science. No text, no extra astronauts, no distorted anatomy, no cartoon appearance, no generic sci-fi city.

Top-down ultra-wide hyper-realistic shot of 4 real people seated at a square dining table. Pull the camera way back so a wide ring of empty floor surrounds the whole scene, giving a clean minimal composition. Scene [ATMOSPHERE + WARDROBE]: the four are styled in [OUTFIT STYLE]. Each one is doing something specific: [DESCRIBE EACH PERSON'S ACTION]. The table itself is a physical prop built to look like an Instagram post viewed from above: - A solid white band runs along the TOP and BOTTOM edges only, with no white on the left or right sides. - Top band: a small drawn profile avatar with [LOGO DESCRIPTION], the handle "[USERNAME]", a tiny blue verified check next to it, and a "..." on the far right. - Bottom band: red heart, comment bubble, and share icons painted on the left, bookmark icon on the right. The center of the table is one flat color: [EXACT COLOR + HEX CODE]. All food, drinks, props, and any hand interacting with them must stay strictly inside that central colored zone, nothing crosses the white bands or the table edge. The floor around the table is [FLOOR COLOR] to create clear visual separation. Professional studio lighting with crisp shadows, 8k, tack-sharp focus.

A playful, high-key studio portrait of [subject] running joyfully across a seamless light gray background, captured mid-stride with one leg lifted and a wide genuine smile. The subject wears a casual oversized outfit with soft neutral tones (or vibrant colors), creating a dynamic sense of motion. Behind them, a simple black hand-drawn cartoon stick figure grabs and stretches the back of their shirt, making the fabric appear elastically pulled as if trying to stop them. The doodle character is integrated naturally into the scene with expressive motion lines and a humorous facial expression. The subject holds a fun prop (such as a dinosaur toy, oversized lollipop, teddy bear, or balloon), enhancing the playful storytelling. Minimalist composition, clean studio lighting, soft shadows, ultra-sharp focus, realistic skin texture, vibrant yet natural colors, whimsical editorial photography, premium children’s fashion campaign aesthetic, highly detailed, photorealistic, 8K.

Rough, gritty anime style with jittery, trembling ink lines and messy, hand-drawn aesthetics. Foreground right third: the weathered, rust-eaten edge of an old fishing vessel deck, with a lone sailor leaning over the railing, hunched and small, staring in awe and terror. Left two-thirds: a colossal cosmic entity rising from the ocean — a massive cephalopod horror in the style of Cthulhu, with enormous writhing tentacles erupting from churning black water, a bulbous alien head with countless eyes, and a body so vast it dwarfs the boat entirely, rendered as a looming blurred silhouette with only glimpses of slimy textured flesh visible through the sea spray. The scale must be overwhelming — the creature's base tentacles alone are wider than the entire ship. The ocean is chaos: violent waves rendered in frantic, sketchy ink strokes. Color palette: deep prussian blue, sickly green-black, and pale moonlight whites, high contrast, no clean gradients, raw and expressive energy.

Turn this photo into a chaotic funny doodle illustration, intentionally messy and low-skill, as if drawn quickly with a cheap marker, crayon, or worn-out felt pen on paper. Create exaggerated facial features with awkward proportions, uneven eyes, oversized head, tiny body, crooked smile, and clumsy anatomy while still keeping the person recognizable. Use rough childish sketch lines, shaky hand-drawn strokes, visible scribbles, overlapping outlines, accidental marks, and random doodles around the scene. Add a simple cartoon-style background with badly drawn buildings, trees, clouds, street elements, and uneven perspective. Coloring should look careless and imperfect, with visible stroke texture, inconsistent fill areas, wax crayon texture, marker bleed, and irregular shading. Include playful imperfections like crossed-out lines, unfinished details, random arrows, tiny notes, stars, swirls, and abstract scribbles. Overall aesthetic should feel humorous, spontaneous, handmade, energetic, goofy, and intentionally unpolished, resembling a child's sketchbook mixed with absurd internet meme art. High texture detail, paper grain visible, asymmetrical composition, awkward framing, expressive doodle chaos, raw sketch energy.

模型能力

grok-image 能做什么

根据提示词生成

提交简洁提示词,通过一个 TTAPI 任务获取图像结果。

常用参数

设置提示词、模型、输出参数和回调即可接入。

异步查询

通过轮询或 Webhook 接收结果,避免耗时生成阻塞界面。

接入产品

Grok Image 的输出可以展示、保存或交给后续流程处理。

计费

按操作计费,单位为 quota

按模型、操作和输出设置扣除 quota。

图像操作官方定价
操作价格计费单位
生成图像2 quota / 张
查询结果免费/ 次请求
API 请求

请求头与提交参数

使用 TTAPI 密钥,按下方字段向接口发送请求。

POSThttps://api.ttapi.io/grok/image

请求头

TT-API-KEYstring必填

填写 TTAPI API 密钥。

Content-Typestring必填

使用 application/json。

请求体

promptstring必填

图像生成提示词。

modelenum选填

默认 grok-imagine-1.0,可选 grok-imagine-1.0、grok-imagine-2.0、grok-imagine-image、grok-imagine-image-quality。

enable_proboolean选填

默认 true。仅 grok-imagine-1.0 和 grok-imagine-2.0 支持,后者必须为 true。

aspect_ratioenum选填

默认 1:1。1.0 支持 1:1、16:9、9:16、3:2 和 2:3;Image 模型还支持 4:3、3:4、2:1、1:2、19.5:9、9:19.5、20:9、9:20 和 auto。

resolutionenum选填

默认 1k。所列模型均支持 1k,grok-imagine-image 和 grok-imagine-image-quality 还支持 2k。

ninteger选填

生成图像数量,范围 1 至 4。

hook_urlstring选填

任务完成时调用的回调 URL。

查看 TTAPI 接入文档
接入示例

发送请求并获取结果

复制当前模型的请求地址、鉴权请求头和请求体。

curl --request POST \
  --url https://api.ttapi.io/grok/image \
  --header "TT-API-KEY: $TTAPI_KEY" \
  --header 'Content-Type: application/json' \
  --data '{"prompt":"A surrealist, dreamlike masterpiece of a lush, misty green hillside enveloped in thick, ethereal fog.","model":"grok-image","enable_pro":true,"aspect_ratio":"16:9","resolution":"1k","n":1,"hook_url":"https://example.com/callback"}'