Prompt-led direction
Describe the subject, action, camera and atmosphere without needing a complex timeline.
Grok Imagine 1.5 Preview + stable model
Choose Grok Imagine 1.5 Preview for prompt-led video with up to seven reference images, or use the stable model for text-to-video with Normal, Fun and Spicy modes.
Output
Your video will appear here
Add a prompt and choose the output settings, then submit a Grok Imagine task.
The models, explained
This page brings together Grok Imagine 1.5 Preview and the stable Grok Imagine text-to-video model. Both turn a written shot description into a short AI video.
Grok Imagine 1.5 Preview accepts up to seven optional reference images. The stable model is prompt-only and adds Normal, Fun and Spicy creative modes; both offer 480p or 720p output.
Every request is submitted as an asynchronous Kie task. The result panel refreshes the latest provider state and stores the finished video in your workspace before it becomes available to preview and download.
Built for fast video ideation
Choose the latest preview workflow or the stable text-to-video model from one page.
Describe the subject, action, camera and atmosphere without needing a complex timeline.
Guide Grok Imagine 1.5 Preview with optional JPG, PNG or WebP reference images.
Use Normal, Fun or Spicy when you select the stable Grok Imagine model.
Follow the latest progress while Kie processes the video and save the final asset to your workspace.
Three simple steps
Choose Preview or stable, describe a specific shot, and let Kie finish the task in the background.
Use 1.5 Preview for optional visual references or the stable model for creative modes.
Write the prompt, add up to seven references for Preview or choose a mode for stable, then set the output.
Wait for the asynchronous task to finish, then preview or download the stored video.
Short videos benefit from prompts that keep the action and camera direction easy to follow.
Questions, answered
Both models create short videos from prompts. Grok Imagine 1.5 Preview can also use reference images, while the stable model offers Normal, Fun and Spicy modes.
Yes. Grok Imagine 1.5 Preview accepts up to seven JPG, PNG or WebP images, each up to 20 MB. The stable text-to-video model does not accept reference images.
Generation time varies with the queue, duration, resolution and provider capacity. The output card continues checking the task until Kie returns the result.
After Kie completes the task, the server downloads the result to private media storage before exposing it through the authenticated output link.
Write a focused prompt and create a Grok Imagine video online through Kie.
Create with Grok Imagine