# Your first fill

URL: https://fulla.dppst.com/docs/generative-fill/

> Select an area, describe what you want and click Generate. The result lands on a new layer, masked by your selection.

Generative fill changes the area you select: add something, change it, or remove it. If you know Photoshop's own generative fill, you already know how Fulla works.

## Make a fill

1. **Select** an area of your photo, with any selection tool.
2. **Describe** in the box at the top of the panel what to add or change there: `a small boat tied to the pier`, `make her smile`. Leave it empty to remove what is selected. Drag the grip under the box to make it taller.
3. Click **Generate**.


A boat added beside a pier with `a small boat tied to the pier`. First try, Nano Banana 2.1, 2K, $0.06.

A card under the box shows the run, with a moving bar and the seconds. Photoshop stays free to use meanwhile, and the button turns into **Cancel**.


## What you get

Each run makes **one new layer**, named after your prompt and masked by your selection, with a soft edge. Your photo itself is never changed, and the whole fill is one undo step. If anything fails, the document is left as it was.

The card shows the model, the size, the time, and the cost of the run and of the session.


Click **Done** once you are happy: the card and the prompt are cleared, and the layer stays.

## Choose a model

The button under the box shows the provider and the model. It opens a searchable list of the models of every provider you connected. **Nano Banana 2.1** is the recommended one. See [Models](https://fulla.dppst.com/docs/models/).


## Next

- [Remove and extend](https://fulla.dppst.com/docs/generative-fill/remove-and-extend/)
- [Reference images](https://fulla.dppst.com/docs/generative-fill/reference-images/)
- [Getting good results](https://fulla.dppst.com/docs/generative-fill/tips/)