Generating in the AI Studio
For creators using My AI → Studio — how a photo run works, what you get back, and what the Video and Voice segments do.
The Studio generates photos. The segmented control also offers Video and Voice, and both are covered at the bottom of this page.
Every run costs one unit from your generation allowance, and every photo is generated as you — the Studio has no way to ask for a generic picture, so your likeness permission and consent have to be in place first.
Generating a photo
Wait for the panel to settle
The Studio asks the platform what it can do the moment it opens. Until that answer arrives the generate button does not act — an empty-looking panel is not necessarily an idle one.
Pick a style chip (optional)
Four looks are offered: Editorial, Studio light, Golden hour and At home. Pick the one closest to what you want, or leave all of them alone — the chip is optional either way.
Describe it under “What should it make?”
Up to 4,000 characters. If the panel offers an “avoid” box, that one takes up to 1,000 characters.
What you get back
Finished images appear in a grid, each one badged AI draft with a spec chip reading 4:5 · PNG and the line “Saved to your Vault.” Every image comes back at 4:5 as a PNG; the screen does not offer a choice of shape or format.
They are ordinary Vault assets from that moment on. You attach them to a post, a message or an order exactly as you would a photo you uploaded — there is no separate AI library to manage.
To remove one, use the bin icon on the draft (Discard this draft). If the file is already attached to a post or delivered on an order, EVVA keeps it and says so: “This file is still in use, so it was kept."
"Another set” and “Use my new prompt”
Two follow-ups appear under a finished set:
- Another set — the same request again, for different images.
- Use my new prompt — the same subject with what you have now typed.
Both are full runs. The note says so — “Each of these runs a new generation.” — and each one costs another unit of allowance. There is no free re-roll.
A run cannot be called back
Once a run has started it runs to the end. A request that has already reached the supplier finishes and is charged, so press Generate photo when you are happy with the prompt rather than to see what happens.
The Video segment
Switch the segmented control to Video and the panel asks for the same things: a style chip if you want one, and a description of what it should make. A finished video is saved to your Vault like a generated photo.
One of the published AI recipes — Bring a photo to life — does this in one step from a single photo, if that is all you need.
The Voice segment
The Voice segment is fully drawn: the reference list switches to your voice samples, the prompt becomes What should it say?, the add tile becomes Record 60s, and the button reads Generate voice note with the note “Voice notes take about 20 seconds.”
Two things about this segment are worth knowing before you start:
- Record 60s registers a reference from your Vault. It opens the Vault picker so you can pick audio you have already uploaded and register it as a voice reference.
- A finished voice note is a Vault asset like any other. Attach it to a conversation from the Vault, the same way you would any other audio file.