Grok
AI Grok (Grok Imagine) prompts covering both image and video generation. Reference them to quickly create sports fan art, comedic composite photos, and game-style action videos.
2026.08.12
把「Grok」做成你想要的样子做一个你自己的「Grok」
用大白话说出来就行——Carat 会为你生成合适的图片和视频。
每天 100 免费积分，现在就开始
Table Tennis Hall Action Portrait
JRPG Marketplace Gameplay Scene
What Are Grok AI Prompts?
These are image and video prompts created with Grok Imagine, the media generation feature built into xAI's Grok app. A single model handles both text-to-image and text-to-video, which makes it common for sports fan art, composite photos, and game-cutscene-style clips.
Where You'll Use It
- Sports fan art and composite photos (player action shots, fan concepts)
- Comedic or meme-style composite images
- Game-cutscene-style action videos
- Quick idea sketches when testing multiple concepts fast
- Casual, high-impact visuals for social media
Why This Tag Is Useful
Grok Imagine prompts generate quickly and handle casual composite or fan-art concepts well, making them useful when you want to test several ideas fast rather than polish one result. Carat runs the Grok family directly. Grok 2 Image calls xAI, and the Grok Imagine line (Image 1.0, Image 2.0 and video) runs through fal, so you can use these prompts on Carat as they are.
Tips for Crafting Prompts
- Put the subject (person or character) and the core action first.
- For composite photos, specify the background and situation clearly, like a stadium or a store.
- For video, include camera work such as a low angle or tracking shot.
- If recreating the look on one of Carat's supported models, lean on detailed scene description rather than Grok-specific shorthand.
Common Issues and Fixes
- Result doesn't look realistic enough: add quality keywords like "photorealistic" or "8k."
- The output resembles a real person too closely: describe a general appearance instead of naming a specific real person to avoid likeness issues.
- Motion in the video looks unnatural: break the action into short, sequential steps.
