Set the visual direction
Generate square images or 16:9 scenes from a text description.
Grok Imagine Image 2.0
Explore visual ideas in square or widescreen format, then use references to evolve the image. A starting point for scene concepts and social visuals.
1K / 2K output
1:1 & 16:9 generation
Up to 3 references
Meet Grok Imagine Image 2.0
Grok Imagine Image 2.0 is from xAI. Explore these workflows and choose generation settings for your project.
Shape the workflow
Choose low or medium quality and compare compositions before refining.
Choose your settings
Edit with up to three references; edited output follows the source image ratio.
Put it to work.
Explore practical workflows and load a prompt to try your own direction.
Social content
Build a visual idea for launches, everyday stories and short campaigns. Start with one strong image and carry the same direction into your next variation.
Storyboards & concepts
Find a film’s visual language with scene concepts and shot studies. Establish the setting and camera angle, then develop each shot with a consistent brief.
Editorial portraits
Explore light, wardrobe and atmosphere for editorial portraits. Define the mood in words, or use a reference to guide the composition.
From an idea to a first result.
Describe the scene
Define the subject, setting, light and composition. For video, add action and camera direction.
Choose your settings
Confirm the model in the generator and choose from its available size, duration or quality options.
Generate and refine
Review the first result, then adjust the prompt or add reference material to develop the idea.
A few things to know
Useful answers before you start creating.
What can I make with Grok Imagine Image 2.0?
Explore visual ideas in square or widescreen format, then use references to evolve the image. A starting point for scene concepts and social visuals.
Which settings are available here?
1K / 2K output; 1:1 & 16:9 generation; Up to 3 references. Available options depend on the variant selected in the generator.
Are these images outputs from this model?
The images come from the corresponding public WaveSpeed model pages, linked below the image. They are not viddeep benchmarks or output guarantees. Results depend on the prompt, references and settings.
What do I need to start?
You can explore the tool and settings first. Submitting a generation requires sign-in and sufficient credits; the generator shows the cost before you submit.
Your next idea starts here.
Bring your brief into the studio.