> For the complete documentation index, see [llms.txt](https://docs.qolaba.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.qolaba.ai/qolaba-2.0/whats-new/image-generation.md).

# Image Generation

Image generation had three separate pages in the old UI: **Text To Image** (generate from a prompt), **Image To Image** (use a reference), and **Image Editing** (inpaint, upscale, remove background, create variations). Switching between these modes required clicking different nav items, losing your place in the process.

All three are now unified in a single **AI Image Generator** workspace. You start by picking a model card or a task chip — Qolaba shows you the right controls for that task automatically. Reference images, editing tools, and generation all coexist in the same canvas.

### Before

![Old Create Images page showing Text To Image active, with model dropdown (Nano Banana 2), select preset, private session toggle, prompt box, Magic Prompt, Keywords, Negative Keywords in the left panel. Right panel shows No. of Generations 1-4, Quality 0.5K-4K, and Dimension grid.](/files/7ce9628d10f8e52a1192478139e46b71c53068c1)

Three separate sidebar items for three modes. Switching from text-to-image to image-to-image meant navigating away and starting fresh.

### After

![New AI Image Generator showing model cards (Flux 1.1 Pro, Gemini Nano Banana, GPT Image, Ideogram) and task chips (Photoreal shot, Illustration & art, Logo & text, From a reference image). Right panel shows Number of images 1-4, Aspect ratio grid, Quality, Advanced.](/files/3e33edcf7ac38b5fb7e780efcce7a9f3dfaaa20f)

One unified canvas. Pick a model card to set your generation style, or a task chip to jump straight to a workflow. The right panel adapts based on what you're doing.

![AI Image Generator canvas showing two generated skincare product images side by side with a glow highlight on the selected one](/files/703b409424688b1b8134bf84ca2a011efa850adc)

Generated images appear directly in the canvas at full size. You can generate multiple variants side by side and compare before downloading.

![Hover state on a generated image showing action icons: expand, wand/edit, thumbs up, thumbs down, download, and a Create video button at the bottom](/files/0dde36d8088ccbf5618a9e054058424c626879fc)

Hover any image to reveal the action row: **Expand** (full-screen view), **Edit** (inpaint, upscale, remove background, variation), **Feedback** (thumbs), **Download**, and **Create video** — which hands the image off directly to the Video Generator.

| Was                                                     | Now — and why it changed                                                                                                                 |
| ------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------- |
| Create Images → **Text To Image** (separate page)       | Left rail → **Image** — the default view is text-to-image                                                                                |
| Create Images → **Image To Image** (separate page)      | Same page → click **＋ From a reference image** chip — upload a reference and the controls adapt automatically                            |
| Create Images → **Image Editing** (separate page)       | Same canvas → hover a generated image → click the **Edit (wand)** icon                                                                   |
| Model dropdown in the left panel                        | **Model cards** in the canvas: Flux 1.1 Pro, Gemini · Nano Banana, GPT Image, Ideogram — with subtitles explaining each model's strength |
| Select Preset, Private Session toggle                   | Still available under **Advanced** in the right panel                                                                                    |
| Magic Prompt, Keywords, Negative Keywords               | **Magic Prompt** button in the prompt bar; Negative prompt under **Advanced**                                                            |
| No. of Generations 1–4, Quality 0.5K–4K, Dimension grid | **Number of images** 1–4, **Quality**, **Aspect ratio** grid — same options, right panel                                                 |