grok-imagine

视频 APIHOT

使用 API 运行 API 文档

Grok Video API 让团队可以用 API 优先的方式，把 Grok Imagine 视频生成接入生产工作流。它在同一个模型家族内支持提示词文生视频、参考图图生视频、视频放大和视频延长。

你可以先在 Playground 中验证提示词、源图片、任务 ID、时长、画面比例和 payload 行为，再把 Grok 视频生成接入应用、创意工具或内容生产管线。

模型类型：

价格

最低成本从 $0.01/s

当前模型

Grok Imagine Text to Video

480p: 2 积分/秒 - 文生视频 - 输出秒数 × 2

720p: 3 积分/秒 - 文生视频 - 输出秒数 × 3

480p: 10 积分/秒 - 文生视频 - 输出秒数 × 10

720p: 14 积分/秒 - 文生视频 - 输出秒数 × 14

美元估算按最佳年付套餐 $0.005/积分计算。最终计费跟随所选 Grok 视频模式和输出设置。

AI 模型

模型版本

提交后生成记录会显示在这里

API 输入

核心请求字段

用可预测 payload 管理提示词、图片 URL、源任务 ID、画面比例和回调设置。

请求 JSON

{
  "model": "grok-imagine/text-to-video",
  "input": {
    "prompt": "A couple of doors open to the right one by one randomly and stay open, to show the inside, each is either a living room, or a kitchen, or a bedroom or an office, with little people living inside.",
    "aspect_ratio": "2:3",
    "mode": "normal",
    "duration": "6",
    "resolution": "480p"
  }
}

使用 Grok Imagine Text to Video 生成

本页的 Playground、请求示例、请求字段和价格都会跟随当前选择的模型类型变化。

查看 API 详细文档

打开完整 API 文档，查看接口、鉴权、任务轮询和 payload 细节。 /models/grok-imagine.md

字段类型说明

promptstringText prompt describing the desired video motion. Required field.

aspect_ratioenumSpecifies the width-to-height ratio of the generated video. Controls the aspect ratio of the output. (2:3, 3:2, 1:1, 16:9, 9:16)

modeenumSpecifies the generation mode affecting the style and intensity of motion. (fun, normal, spicy)

durationnumberThe duration of the generated video (in seconds) (6-30). (Minimum: 6, Maximum: 30, Step: 1)

resolutionenumThe resolution of the generated video. (480p, 720p)

nsfw_checkerbooleanDefaults to false. You can set it to false based on your needs. If set to false, our content filtering will be disabled, and all results will be returned directly by the model itself.

能力

Grok Video API 为视频工作流带来的能力

页面按 API 优先的结构组织：先测试模型，再理解字段，最后进入集成。

文生视频生成

把详细提示词转成短视频概念，适合社媒内容、广告、产品瞬间和快速创意探索。

图生视频运动

基于源图片生成动态视频，同时保留团队已经确认的主体、构图和视觉方向。

放大和延长

对成功的 Grok 短片进行视频放大，或通过延长任务继续已有生成结果。

API 就绪 payload

在一条集成路径中处理模型 ID、提示词、图片 URL、任务 ID、画面比例、回调和任务轮询。

快速开始

从 Grok 视频测试到生产请求

适合团队先验证输出质量，再把 Grok 视频生成接入自己的应用。

1. 选择模式

提示词直出选择文生视频；需要源图动起来选择图生视频；已有结果再进入放大或延长模式。

2. 提交生成

使用账户密钥、模型 ID、提示词、可选图片 URL 或源任务 ID，以及回调地址调用 API。

3. 轮询并复用

跟踪任务直到成功，然后把输出 URL 用于产品、营销工作流或下一轮视频任务。

使用场景

你可以用 Grok Video API 创作什么

当产品需要可重复的 Grok 视频输出，而不是一次性手动生成时，可以直接接入 API。

TikTok、Reels、Shorts 的提示词短视频

商品静图动效和活动视觉

快速分镜和概念视频探索

广告、缩略图和落地页视频变体

成功生成短片的高清放大版本

面向迭代创意管线的视频延长

社区案例

社区公开的 Grok 视频成功案例

查看数据库中公开且已成功生成的 Grok 视频作品，打开案例可继续查看提示词、参数和生成结果。

videoGrok Imagine

Real life scene, bank Holiday monday, new build living room on sofa, randy dominant 70-year-old Salford bloke next door dad wearing a crisp diffferent cotton business shirt with a sheen large stiff pinched collar, top buttons undone, tucked tightly in

Real life scene, bank Holiday monday, new build living room on sofa, randy dominant 70-year-old Salford bloke next door dad wearing a crisp diffferent cotton business shirt with a sheen large stiff pinched collar, top buttons undone, tucked tightly into dark smart trousers with a black belt., he is otk disciplining a 40 year old, 5 foot, thug scally bloke, who is wearing simlar, filthy talking, swearing

2026年7月12日查看详情

videoGrok Imagine

Use the uploaded image as the only reference. Create an 8-second vertical (9:16) ultra-realistic cinematic fashion video. The woman must remain exactly the same person as in the reference image. Preserve her facial features, skin tone, hairstyle, body p

Use the uploaded image as the only reference. Create an 8-second vertical (9:16) ultra-realistic cinematic fashion video. The woman must remain exactly the same person as in the reference image. Preserve her facial features, skin tone, hairstyle, body proportions, natural expression, and overall appearance consistently throughout the entire video. The dress must remain exactly the same dress shown in the reference image. Do not redesign, restyle, or modify it in any way. Preserve the exact yellow color, the same cotton fabric texture, embroidery, covered buttons, stitching, round neckline, sleeves, waist seam, silhouette, length, wrinkles, folds, and every visible detail exactly as in the original image. The back of the dress is extremely important. Keep it exactly the same as the reference image. It has a simple, clean back with only the waist seam. There must be NO belt, NO sash, NO ribbon, NO bow, NO back tie, NO pleats, NO gathers, NO decorative elements, NO extra seams, and NO redesign of the back. The woman begins sitting naturally on the sofa in the seaside lounge. She smiles gently, stands up naturally, then slowly walks through the elegant Mediterranean seaside corridor beside the ocean. The corridor has white walls, blue wooden railings, Mediterranean flowers, green vines, and beautiful coastal plants. The sea remains visible beside her during the walk. Her walking is slow, elegant, and completely natural. She has realistic posture, subtle arm movement, natural breathing, and soft hair movement from the sea breeze. The dress moves only with realistic fabric physics while keeping its exact original shape and design. The camera starts with a medium shot, then smoothly follows behind her with a stabilized cinematic tracking shot, clearly showing the original back of the dress while she walks. At the end of the corridor, she gently stops, turns her head toward the sea, and smiles softly while looking at the horizon. Bright Mediterranean afternoon sunlight, realistic shadows, natural ocean reflections, crystal-blue water, gentle waves, subtle movement of flowers and leaves in the breeze. Professional luxury fashion commercial, photorealistic, 4K HDR, full-frame cinema camera, natural skin texture, sharp details, stable camera movement, no AI look. Negative Prompt: No dress redesign, no color change, no fabric change, no belt, no sash, no ribbon, no bow, no pleats, no gathers, no extra seams, no accessories, no body reshaping, no face changes, no identity drift, no clothing deformation, no flickering, no warped hands, no distorted anatomy, no AI artifacts.

2026年7月12日查看详情