Text-to-Image
Write the prompt first, then choose quality and size settings, and generate multiple image versions you can refine from the gallery.
Text-to-Image is a prompt-first generation page. You can write your own prompt from scratch or upload a reference image to extract a prompt first, then switch between Standard, Enhanced, and HD modes to produce multiple creative directions.
When To Use
Use this page when you know what angle or value proposition you want to show, but you do not yet have the right asset.
It is useful for hero-image exploration, ad-creative drafts, scene concepts, and style testing.
Workflow
The page is structured around three parts: prompt input, parameters, and a reusable gallery.
- 1
Step 1: Pick the Quality Mode and Write the Prompt
Start by choosing Standard, Enhanced, or HD mode. Then describe the subject, scene, lighting, and selling-point expression in the Prompt area.
- 2
Step 2: Set Size, Resolution, and Output Count
Standard mode focuses on size. Enhanced and HD modes unlock more quality-related parameters such as resolution. The footer lets you choose how many versions to generate in one run.
- 3
Step 3: Review the Gallery and Edit Again
Finished outputs go into the gallery, where you can refresh task status, download originals, edit again, or delete weak results.
Key Controls
The biggest quality differences usually come from the prompt and the quality settings, not from minor button choices.
- Prompt: defines the subject, action, scene, and style
- Quality Mode: balances speed, detail, and cost
- Size / Image Size: controls aspect ratio and output sharpness
- Generation Count: creates multiple variants for side-by-side comparison
Best Practices
Use a lower-cost mode to explore directions first, then move the best prompt into Enhanced or HD mode for production-ready outputs.
If a result is close, do not start over. Reuse Edit Again from the gallery and make small changes around the working direction.
- Describe the subject first, then scene, lighting, and camera feel
- Generate a small number of versions per round so comparison stays clear
- Save winning prompts so the team can reuse them later