Beta Not the live site. Pages in review Live site →
Loft

Learn/Get started

Ask for an image

Your first generation: ask the Assistant for a new image of your character, in plain words. Follow along in the project from the last video.

Follow along

Step by step

STEP 01

Your tokens

Your balance reads 250: the tokens a new organization gets to try Loft, enough for these videos. They last two days. Open Characters, then Wren.

Your tokens
STEP 02

Open the Assistant

Click Assistant. On Wren's page, it knows you're talking about Wren.

Open the Assistant
STEP 03

Ask in your own words

Say what you want, the way you'd brief an artist: Make a turnaround of Wren: front, side and back, full length, on a plain background.

Ask in your own words
STEP 04

It picks the model and the reference

Press Enter. The Assistant chooses a model and uses Wren's published design as the reference, so the new image keeps her look. It says what it did, and that the image is on its way.

It picks the model and the reference
STEP 05

It lands on Wren

In a minute or so the image lands in Wren's files, marked New, and a note pops up to say it's done. Yours will look different from this one.

It lands on Wren
STEP 06

Look at it

Close the Assistant, and click the new image to see it full size.

Look at it
STEP 07

Keep it: Selected

If you like it, press 2 to mark it Selected: your pick, for now. Publishing comes later, once you've decided. Not quite right? The next video directs it.

Keep it: Selected
STEP 08

The same, by hand

Close the viewer. Everything the Assistant did, you can do yourself on the Generate tab: pick a model, write the prompt, add images. The model it used is at the top, under Recently used. And the balance is a little lower: that's what the image cost.

The same, by hand