Ask Orsola to create an image or a video

Ask for one

Describe what you want in a conversation and say you want a picture:

"Make me an image of a quiet home office at dawn, warm light through a window, plants on the desk, no people. Wide, for a presentation background."

Video works the same way, and takes longer to produce.

Describe it so the result is usable

The difference between a usable image and a near miss is nearly always the description. Worth saying:

• The subject, specifically.
• The mood and lighting — warm, stark, overcast, evening.
• The style — a photograph, a flat illustration, a line drawing.
• The shape — wide for a banner, square for a social post.
• What must not be in it. "No people", "no text" — exclusions do a lot of work.

Text inside generated images is unreliable across every tool of this kind. If you need words on it, add them yourself afterwards rather than asking for them.

Refine from what you got

Treat the first result as a draft: "same but colder light and no plants", "more space at the top for a headline". Adjusting beats re-describing from scratch, because everything you liked stays.

If it is a long way off, change the description rather than repeating it — the same words tend to produce the same kind of result.

What it costs

Making images and video is one of the most credit-hungry things Orsola does, and video considerably more than images. Both show under Media Generation in check your balance and usage.

Ten refinements is ten generations. Worth a little thought about the description before the first one rather than iterating your way through a month's credits.

Finding them afterwards

Everything generated goes to your gallery. Filter to Images or Videos and sort by date — usually faster than finding the conversation it came from.

Where to go next

What credits are, and what uses them if generating a lot, and ask Orsola to make a document or spreadsheet for written output.