Hey everyone,
Got a prompt trick that always works for you?
We’re building a list of TIPS and TRICKS for image generation using OpenAI tools.
The goal: help new users avoid wasting prompts and time by learning from what we already know.
If you’ve spent time figuring out how to write better prompts, this is your chance to pass that on. Just drop your tips below.
Why we’re doing this:
- New users often don’t know what works and what doesn’t.
- Good image generation takes trial and error—but we can reduce the “error” part.
- Sharing real examples and techniques will help everyone get better results faster (and maybe they will use fewer prompts if they get the desired result).
What we’re looking for:
Prompt structure tips (e.g., how to be specific, use references, modifiers)
Words or styles that improve results (e.g., “cinematic lighting,” “ultra detailed”)
What NOT to do (common mistakes or bad habits)
Tips for specific styles: photo, illustration, cartoon, surreal, etc.
Any tool/workflow tricks (image upscalers, editing, prompt testing, etc.)
Even if you only have one small tip, post it—it might be a game-changer for someone else.
Reply below with anything you’ve learned. Let’s make it easier for the next person.
Example to kick things off: Using terms like “backlighting” or “golden hour” can dramatically improve the lighting in portrait prompts.