The 30-second workflow
Open the studio, upload a portrait, pick a visual direction, generate. That is the entire loop, and it finishes in seconds. Everything below is about doing those four steps deliberately, because each one has a quality lever hidden in it.
Choosing the visual direction
Three presets cover most needs. The editorial direction sharpens light and pushes a stylized, magazine-cover feel. The dream direction softens everything with gentle color grading. High fidelity is the conservative choice: it keeps the expression and proportions closest to the original photo.
When in doubt, generate with high fidelity first. It gives you a baseline to judge whether a problem came from the photo or from the styling.
Settings that actually matter
Output resolution is the one setting worth caring about: 1K is fine for chat apps, 2K holds up on feeds and re-compression. Anything else in the panel you can leave on defaults until you have a specific reason to change it.
Free generations are public by default; once you sign in you control visibility per generation.
From output to post
Download the HD file rather than screenshotting it. Screenshots compress, and the compression artifacts are exactly what makes a swap look fake when it is re-shared. Post the clean file, keep captions short, and let the image do the work.
The full loop, shoot, swap, caption, post, takes under five minutes once you know your source photos are good. That speed is the whole point: the faster the loop, the more shots you take at the trend.