Generating images

Mavic can generate brand-consistent images for social posts, blog headers, product variations, and ad creative. This article covers the image-generation workflow, how to get sharp results, and how to apply your Brand Kit and Templates.

How to start

You have three entry points:

1. From a chat prompt

Generate a square image of a coffee cup on a wooden table, soft natural light, minimalist, in @Brand Colors.

Mavic generates the image inline. Refine via follow-up prompts.

2. From the Image Generation workflow

Click the workflow icon in the chat composer and pick AI Image Generation (or Image Generation → Brand Image etc., depending on workspace). The form asks:

• Subject — what's in the image
• Style — photorealistic, illustration, 3D, cinematic, anime, custom
• Mood / lighting — golden hour, studio, moody, bright
• Composition / framing — close-up, wide-angle, top-down
• Brand application — apply your Brand Kit colors? Logo overlay?
• Aspect ratio — square (1:1), portrait (4:5, 9:16), landscape (16:9)

Mavic builds a polished prompt from your inputs and generates the image.

3. Product image variations

Inside Brand DNA → Products & Services, open a product and click Generate variations. Mavic produces alternative shots (different backgrounds, contexts, angles) using the original product photo as a reference.

Applying your brand

The cleanest way to keep images on-brand:

• @-tag a Template in your prompt — Mavic uses it as the layout
• @-tag your Logo to overlay it
• Reference your Brand Kit — e.g. "Apply our brand colors as the background"
• Use the workflow's "Apply brand" toggles — quicker than typing it out

Photorealistic vs other styles

Mavic defaults to photorealistic outputs. To get a different style, say so:

"Style: flat illustration, bold outlines, vibrant brand colors."

>

"Style: anime, soft pastels, dreamy lighting."

>

"Style: cinematic photo, shallow depth of field, golden hour."

Be specific. Vague terms like "futuristic" produce uneven results; "neon-lit cyberpunk metropolis with towering skyscrapers" produces consistent results.

Refining a generated image

After Mavic produces an image, you can:

• Type follow-up changes in the Add feedback box: "Make the lighting warmer and reduce the clutter on the right."
• Generate variations from the same prompt
• Restore a previous version
• Compare versions side-by-side

See Editing & versioning generated assets.

Tips for sharper output

• Describe the camera and lens — "shot on 50mm portrait lens, shallow depth of field"
• Describe the lighting — "soft window light from the left, gentle shadows"
• Avoid contradictions — don't ask for "minimalist and busy" in the same prompt
• Iterate rather than restart — small changes preserve what's working
• Use templates for predictability — they remove most variability

Aspect ratios and where to use them

Image attributes Mavic infers

When you don't specify, Mavic chooses sensible defaults:

• Lens: 50mm for portraits, wide-angle for landscapes, macro for close-ups
• Depth of field: shallow for portraits, deep for environments
• Lighting: realistic with soft shadows and atmospheric depth

Override any of these in your prompt if you have a specific look in mind.

Multi-image consistency

When you ask for multiple images in one go, Mavic tries to maintain a cohesive visual identity — same color palette, lighting style, and subject treatment. If you need extra consistency, mention it explicitly:

"Generate three images of the same product on a beach, sunrise, blue overcast, and dusk. Keep the product position and angle identical across all three."

Where generated images go

Every image lands in Library → Generated Images. From there you can download, reuse in posts, or @-tag in future prompts.

Related articles

• Templates
• Brand Kit
• Editing & versioning generated assets
• Your library

Updated: May 2026

Updated on: 07/05/2026