Custom AI Image Generator — Full Creative Control with Brand Context
When none of the pre-built workflows match your vision, Custom Image gives you the full power of Nano Banana 2 — the most capable multi-reference AI image model — with your brand context automatically applied. Write any prompt, add up to 14 reference images, generate anything.
At a glance
What this workflow does
AI Model
Google Nano Banana 2
Image generation — 3 credits per image
Output
Image
1:1, 4:5, 9:16, 16:9 aspect ratios
Scale
Single generation
3 credits per run, instant output
Example output
Here's what you can generate
Real outputs from this workflow. Supply your brand colors, product description, and optional reference image.
How it works
Three steps. That's it.
Pick this workflow
Select Custom Image from the workflow catalog. It's pre-tuned for brand & identity content.
Add brand reference
Add your brand name, accent color, and an optional reference image. socialAF auto-injects your context into the prompt.
One-click generate
Hit generate. Google Nano Banana 2 runs the workflow and delivers your image — ready to publish.
Under the hood
Models powering this workflow
Google Nano Banana 2
PrimaryImage generation model — 3 credits per image. Handles photorealistic detail, brand color integration, and composition.
Flux 2 Pro
Automatic fallbackActivates automatically if the primary model is unavailable or rate-limited — ensuring your generation always completes without manual intervention.
Best for
When to use Custom Image
Power users and prompt engineers who want full control over AI image generation
Agencies creating highly customized brand visuals that don't fit standard templates
Brands testing creative directions before committing to a workflow pattern
Developers and content teams using the REST API for programmatic image generation
Creators exploring new visual styles by combining multiple reference image sources
FAQ
Frequently asked questions
What AI model does the Custom Image workflow use?
Nano Banana 2 is the primary model — it is one of the most capable image generation models available, supporting up to 14 reference images simultaneously. This makes it ideal for complex brand-context generations where you need the AI to understand your specific product, style, and aesthetic. Flux 2 Pro is the automatic fallback.
How many reference images can I use?
Up to 14 reference images can be uploaded for a single Custom Image generation. Use multiple references to provide richer brand context: product photos, color swatches, style inspiration, competitor images, and mood boards — all processed together.
What prompt techniques work best with Custom Image?
Be specific about subject, lighting, composition, and mood. Specify the camera style ('product photography', 'editorial portrait', 'flat lay') and the output use ('Instagram post', 'LinkedIn banner'). The model auto-injects your brand name and color — you don't need to add those.
What aspect ratios can I generate with Custom Image?
Custom Image supports all four major aspect ratios: 1:1 (Instagram square), 4:5 (Instagram portrait), 9:16 (Stories/TikTok), and 16:9 (YouTube/LinkedIn). Multi-ratio generation produces all four in a single run.
When should I use Custom Image instead of a pre-built workflow?
Use Custom Image when: you need a very specific scene that no pre-built workflow covers, you are experimenting with prompt variations, you need to incorporate many reference images, or you are a power user who prefers writing prompts directly. Pre-built workflows are faster and produce more consistent results for their specific use cases.
More from Brand & Identity
Related workflows
Start creating with Custom Image
52 pre-built workflows, AI content generation, brand voice training, and autopilot publishing. $20/mo — cancel anytime.