Shapes for every screen
Thirteen aspect ratios cover square posts, vertical stories, tall phone wallpapers and wide banners, so you can generate in the shape you will publish instead of cropping later.
Grok Imagine by xAI · on MossImage
Grok Imagineimage generator.
Turn a concrete scene into a still image with Grok Imagine Image 2.0 by xAI. Pick from 13 aspect ratios, including tall phone formats, choose Standard or Medium quality, and add up to three reference images.
Try:
What you can make with Grok Imagine
An imaginative scene still needs a clear structure. Describe the main subject, its size and position, then the setting and the light. If you combine unusual things, explain how they relate: “a tiny greenhouse on a desk beside a reading lamp” tells the model more than “a surreal creative masterpiece”. Change one thing at a time when you iterate, so you can see what each change did.
The model covers square, portrait and landscape shapes, plus the tall 9:19.5 and 9:20 formats of modern phone screens and wide 20:9 and 2:1 banners. Add up to three references when a face you may use or a product must match. For edits limited to boxed areas, the editor’s One area mode runs on GPT Image 2.5 and tells you before it runs.
Strengths and controls
Start concrete, then vary
Write who is where, doing what, from which angle. Once a scene works, change only the background, the light or the outfit in the next run to compare results fairly.
Standard or Medium quality
Choose Standard for quick drafts and Medium for the final image, at 1K or 2K. The editor shows the cost of each combination before you run.
Settings on MossImage
Straight from the MossImage model catalog. Credits are per image; the editor shows the exact total for your settings before you confirm.
| Version | Can do | References | Sizes | Quality | Aspect ratios | Credits per image |
|---|---|---|---|---|---|---|
| Grok Imagine Image 2.0 | Generate · Edit whole image | Up to 3 | 1K, 2K | Standard, Medium | 1:1, 3:4, 4:3, 9:16, 16:9, 2:3, 3:2, 1:2, 2:1, 19.5:9, 9:19.5, 20:9, 9:20 | 1K 12–18 · 2K 19–25 |
Ways sellers and creators use it
Social posts and stories
Pick the shape of the platform first, 3:4 for feeds or 9:16 for stories, and leave room for captions and stickers.
Concept scenes
Sketch an idea for a campaign or a channel banner: an unusual object in an everyday place, described clearly enough to read at a glance.
Character portraits
Describe a fictional person, their clothing, the setting and the light. Add a reference image you are allowed to use if the look must match.
How it works on MossImage
- 1
Write a focused brief
Start with one subject and a clear composition. With reference images, say what must stay and what may change. The templates below are a good starting point.
- 2
Pick the version and settings
Choose the version, shape, size and quality above. The Generate button shows the credits before anything runs.
- 3
Review, then fix one area
Check text, faces, hands and product details. To change one part, open the image, choose One area, box it and write what should change; everything outside the box stays as it is.
Prompts to try
Café portrait
A fictional young woman reading by a café window, soft morning side light, an open book and a cup of tea on the wooden table. Candid, natural colours, waist-up vertical portrait.
Person, action, place and light in one sentence each.
Use this promptConcept scene
A tiny glass greenhouse full of herbs on a cluttered wooden desk, next to a brass reading lamp and a stack of notebooks. Warm evening light, shallow depth of field, cosy and detailed.
Explain how the unusual parts fit together physically.
Use this promptIdeas from the MossImage gallery
We have not published Grok Imagine examples yet. These gallery images were made with other models, named on each card; use their prompts as a starting point.
Before you generate
- Check text, faces, hands, logos and product details on every result before you publish it.
- Only upload photos you have the right to use, and say clearly what must stay the same.
- Settings come from the menus, not the prompt: writing “4K” in a prompt does not change the output size.
- Uploads and results are private to your account and not indexed. Failed jobs are refunded automatically.
Grok Imagine questions
Which Grok model does MossImage use?
Image 2.0 by xAI, the still-image model in the Grok Imagine family. Every image you make is labelled with the model that made it.
How many reference images can I add?
Up to three. Use them for the details that must match, such as a face you have permission to use or a product, and describe everything else in text.
Can I edit only part of a photo with Grok?
Grok edits the whole image from your instruction. To change only boxed areas and keep the rest pixel-identical, use One area in the editor, which runs on GPT Image 2.5.
Can I make videos with Grok Imagine here?
No. MossImage makes still images only. This page covers the image model; xAI’s video models are not offered on MossImage.
How much does Grok Imagine cost?
Standard costs 12 credits at 1K and 19 at 2K; Medium costs 18 at 1K and 25 at 2K. The total is shown before you confirm, and failed jobs are refunded.