grok-imagine-video-1.5

由 xAI 提供视频

经非官方渠道用文字、参考素材、首尾帧或带时间的关键帧生成 1 至 15 秒视频,最高 1080P。

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 quota480P 或 720P
视频示例16:9 landscape 预览
视频示例

15-second cinematic historical period-drama scene, realistic, intimate, elegant, restrained, filled with quiet longing. A woman remains SEATED at a writing desk beside a rainy leaded window, writing a personal letter to someone she once loved. Aristocratic interior, aged wood, period dress, feather pen, old paper. Warm flickering candlelight contrasts with cool blue-gray rainy daylight. Emotional rhythm: WRITE → THINK → WINDOW GLANCE → WRITE → PAUSE → SECOND WINDOW GLANCE → FINISH → LEAN BACK → LOOK OUT WINDOW. 0–3s: Close-up of feather pen slowly writing beside a candle. The pen briefly hesitates above the paper. Camera gently reveals her thoughtful face. 3–6s: Writing slows. She pauses, subtly tightens her lips, then makes a brief FIRST glance toward the rainy window. She returns to the page and continues writing. 6–9s: The pen pauses again. Her free hand lightly touches the letter. She makes a SECOND, longer window glance. Her eyes soften and lose focus slightly, as if remembering someone far away. Candlelight warms one side of her face while cool window light shapes the other. 9–12s: She returns to the page and slowly writes a final line. Her expression softens from emotional weight into tenderness and longing. 12–15s: She stops writing, lowers the pen, and gently LEANS BACK INTO THE CHAIR. She remains seated. Her shoulders relax as she looks toward the rainy window with a soft, distant gaze and the faintest private smile. End there. Performance: Natural micro-expressions: small writing hesitations, hovering pen, trembling lashes, subtle breathing, slight mouth tension, softened eyes, brief loss of focus. No crying or melodrama. Camera: One continuous smooth cinematic shot. Start close on pen, paper and candle, drift naturally to hands and face, capture both window glances, then end on her seated profile looking toward the window. No abrupt cuts or flashy movement. Audio: No dialogue or voiceover. Soft pen on paper, rain, candle ambience, fabric movement, light wind against glass, subtle chair creak. Minimal melancholic period piano or strings, warming slightly near the end. STRICT: She stays seated for the entire scene. Never stands or walks to the window. Two window glances while writing. Writing is repeatedly interrupted by thought. End leaning back in the chair looking toward the window. No subtitles, captions or on-screen text.

图生视频 · 非官方渠道 · 15s · 16:9 landscape
生成示例

从提示词到成片

模型能力

grok-imagine-video-1.5 能做什么

根据提示词生成

提交简洁提示词,通过一个 TTAPI 任务获取视频结果。

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-imagine-video-1.5 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。

非官方渠道按秒计费官方定价
分辨率价格计费单位
480P / 720P2 quota / 秒
1080P4 quota / 秒
API 请求

grok-imagine-video-1.5 操作和请求字段

使用 TTAPI 密钥,按下方字段向接口发送请求。

生成POST
/grok/generations

使用 video_length、resolution_name、参考图和可选回调提交非官方 Grok 视频任务。

查看接口文档
首帧POST
/grok/generations

在 1.5 Fast 或 1.5 中使用 first_frame 指定开场图像。

查看接口文档
尾帧POST
/grok/generations

在 1.5 Fast 或 1.5 中使用 last_frame 指定结尾图像。

查看接口文档
首尾帧POST
/grok/generations

给 1.5 Fast 或 1.5 指定开场和结尾图像。

查看接口文档
关键帧POST
/grok/generations

在 1.5 Fast 或 1.5 片段中的指定时刻固定最多 4 张关键帧。

查看接口文档
POSThttps://api.ttapi.io/grok/generations

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

生成提示词。基础版、1.5 和 Lite 最多支持 4,096 字符,1.5 Fast 支持更长提示词。

modelenum选填

默认 grok-imagine-video-1.5-fast,也支持 grok-imagine-video、grok-imagine-video-1.5 和 grok-imagine-video-1.5-lite。

aspect_ratioenum选填

默认 16:9,支持 2:3、3:2、1:1、9:16 和 16:9。

video_lengthstring选填

默认 10 秒。Fast 支持 6 至 30 秒,基础版、1.5 和 Lite 支持 1 至 15 秒。基础版用单图时最长 15 秒,多图时最长 10 秒。

resolution_nameenum选填

默认 720p。所有模型均支持 480p 和 720p,1.5 和 Lite 支持 1080p,Lite 的 1080P 由 720P 放大。

refer_imagesstring[]按条件填写

参考图 URL。Fast 最多支持 7 张,基础版支持多张,1.5 的参考图工作流仅用一张,其他 1.5 工作流使用文本或显式首尾帧设置。

first_framestring选填

首帧图 URL,1.5 Fast、1.5 和 Lite 的首帧工作流支持。

last_framestring选填

尾帧图 URL,1.5 Fast 和 1.5 支持。first_frame 和 last_frame 使用同一图像可形成循环效果。

reference_audiosobject[]选填

用于旁白或配音的角色对象,每项含 voice_id,最多 3 项。

hook_urlstring选填

任务完成或失败时调用的 URL,也可通过 Fetch 查询结果。

keyframesobject[]按条件填写

1.5 Fast 或 1.5 的关键帧生视频必填。最多 4 项,每项含 image: { url } 和 timestamp_s。时间戳须位于片段内,相隔至少三分之一秒。可与 first_frame 和 last_frame 搭配。

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/generations \
  --header "TT-API-KEY: $TTAPI_KEY" \
  --header 'Content-Type: application/json' \
  --data '{"model":"grok-imagine-video-1.5","prompt":"a young man closes his eyes on a crowded nightclub dance floor","aspect_ratio":"2:3","video_length":"6","resolution_name":"1080p","refer_images":["https://example.com/first-frame.jpg"],"reference_audios":[{"voice_id":"luna"}]}'