Introduction
An introduction to Image Generation in Qolaba — capabilities, available models, interface overview, and how to generate your first image.
Qolaba's Image Generation workspace gives you access to 15+ AI image models in one interface — from Nano Banana and GPT Image to Flux, Stable Diffusion, Ideogram, Recraft, and more. Generate images from text prompts, transform existing images using references, edit specific areas, remove backgrounds, upscale for production, and create variations — all with credit costs shown before you generate.
What You Can Do
Text-to-Image : Generate images from written prompts. Describe a subject, environment, style, and mood — and the model produces it. Apply presets for consistent styling, use keywords for precise visual control, and adjust quality and dimensions to match your output format.
Image-to-Image : Upload one or more reference images to guide generation. The model uses your references to influence composition, style, or subject — producing outputs that stay visually consistent with your input. Useful for brand asset generation, product visualization, and style-consistent creative work.
Image Editing : Edit, enhance, and transform existing images using four dedicated tools:
Inpainting — selectively replace or modify specific areas without regenerating the full image
Upscaling — increase resolution 2x to 4x for print and high-fidelity output
Background Removal — cleanly isolate subjects from their backgrounds
Image Variation — generate creative variations inspired by reference images
Who It's For
Marketing & Creative Teams — Ad creatives, campaign visuals, social media content, brand assets
Designers — Concept visualization, mood boards, style exploration, production-ready assets
Product Teams — Product mockups, UI concept illustrations, feature visuals
Content Creators — Thumbnails, cover art, visual storytelling, editorial imagery
Agencies — Multi-client creative production with workspace-level organization and privacy controls
How to Access
Go to the left navigation panel
Click Image Generation
This opens the dedicated Image Generation workspace.
Interface Overview
The Image Generation workspace is organized into two areas:
Input Controls : Where you configure everything that goes into a generation — model selection, presets, session visibility, prompt, Magic Prompt, keywords, and reference image uploads.
Output Controls : Where you define the technical parameters — number of generations, quality, and dimensions. Credit cost updates as you adjust settings and is displayed before you generate.
Available Models
Qolaba provides access to 15+ image generation models across different providers. Each model has distinct strengths, reference image support, and credit cost:
Nano Banana Pro
49
Realistic visuals, polished creatives, high-quality renders
Nano Banana 2
25
Balanced quality and cost, general-purpose generation
GPT Image 2
18
Concept art, structured scenes, strong prompt comprehension
Seedream 4.5
21
Stylized visuals, artistic and expressive images
ImageGen Fast
3
Most cost-effective, rapid iterations
Credit cost per image varies by model. Total cost = credits per image × number of generations. The exact cost is displayed before you generate. See Image Models → for a full breakdown of model capabilities and recommendations.
Generate Your First Image
Step 1 — Open the Image Generation workspace from the left navigation panel.
Step 2 — Select a model. If you are unsure where to start, Nano Banana 2 is the recommended general-purpose model for most generation tasks.
Step 3 — Select a preset for consistent styling, or leave it unset to work from your prompt alone.
Step 4 — Set your session visibility — Private for client or proprietary work, Public to share outputs to the community gallery.
Step 5 — Write your prompt. Be specific about subject, environment, style, and mood. Use Magic Prompt to enhance it automatically, and add keywords for precise visual direction.
Step 6 — Configure your output — quality, dimensions, and number of generations.
Step 7 — Review the credit cost, then click Generate.
Start at 1K quality with a single generation when testing a new prompt or model. Validate the direction before upgrading to higher quality or running a batch — this is the most credit-efficient way to iterate.
What's in This Section
Last updated