Search for a command to run...
Creating and editing images.
Open the Image Generator and pick a model from the sidebar. Here's a quick guide:
— FLUX 2 Pro: Best all-round model for quality and prompt adherence.
— Seedream 4.5: Excels at photorealism and faces.
— Nano Banana: Fastest and cheapest — great for quick drafts and iteration.
— GPT Image 1.5: Strong at text rendering and complex compositions.
Each model shows its OINKS cost per image before you generate.
Be specific about what you want: describe the subject, style, lighting, camera angle, and mood. Use the built-in Prompt Enhancer (the sparkle button above the prompt field) to automatically improve your prompt.
You can also browse Community posts to see prompts that produced great results.
Several models support image-to-image generation. Upload a starting image and describe the changes you want.
For character consistency across multiple images, try the Character Studio to train a LoRA on your character, then use it in any generation.
Generated images are typically 1024x1024. To increase resolution, use the Image Upscaler powered by Topaz AI — it can scale up to 4K.
To edit specific regions of an image, use the Inpainting tool — paint over the area you want changed and describe the replacement.