Start here
Getting Started
Use this guide when you are opening PromptAnt for the first time and want a clean path from idea to saved result. It covers the minimum decisions you need before the first run, then shows how to keep useful outputs available for the next iteration.
At a glance
- Pick the media type first so the composer opens with the right model controls.
- Use one focused prompt goal, then add references only when they clarify the visual target.
- Treat strong outputs as reusable assets instead of one-off screenshots.
Step 01
Create your first result
Start from AI Image or AI Video, then describe the subject, style, framing, and any constraints that matter. Short prompts work for exploration; more structured prompts are better when you need repeatable results.
Before submitting, check the selected mode and uploaded references. The task list keeps recent runs visible, so you can compare outputs without losing the prompt that produced them.
Step 02
Use references to guide the model
Reference images are useful when the scene, product, character, or composition needs to stay recognizable. Upload only the images that directly support the current task.
When moving from image exploration into video, reuse the strongest still as a visual anchor and keep the motion request specific, such as camera movement, subject action, or atmosphere.
Step 03
Keep useful results organized
Promising generations can be opened from their detail page, downloaded, copied, or reused as starting material for another prompt. This keeps your workflow from depending on scattered screenshots.
Use Favorites and Assets as your working library. They are designed for repeated comparison, later remixing, and collecting ideas that are not finished yet.