> ## Documentation Index
> Fetch the complete documentation index at: https://docs.goosybear.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Making images

> The images lane: the four one-click starts, working a board, iterating on a single image, and what happens when an image is finished.

**Images** in the side menu is where pictures get made — thumbnails, post
images, ad concepts, anything else you need a picture for. Every image is made
on a [board](/work/boards), so nothing you try is ever lost and nothing has to
be right first time.

## Starting something

The lane opens on **What are we making today?** — four one-click starts:

| Start                               | What arrives                 |
| ----------------------------------- | ---------------------------- |
| **Thumbnail for your latest video** | 3 concepts · 16:9            |
| **This week's post images**         | 5 images · 4:5               |
| **Your brand in 6 styles**          | 6 images · square            |
| **Ad concepts from a product shot** | 4 concepts · bring one photo |

Pressing one creates a board with a matching name and drops you into it with
the request already typed. **Nothing runs until you press send** — the start
fills the box, you decide whether to change it first.

If none of the four is what you want, **New board** in the header gives you an
empty one and you ask for whatever you like.

Under the starts, **Jump back in** shows the boards you were last working in,
each led by what is on it. **All boards** opens the full list, and the count
beside the page title tells you how many you have.

## Asking for images

Inside a board, the conversation runs down the left and what you make lands on
the canvas. Ask in your own words — *"six thumbnail options, bold type, room
for a face on the right"* — and the images arrive on the board.

Above the box are the settings the next run uses. Each one is a button you
press to change:

* **Size** — the shape the images come out in. Left alone it follows what you
  asked for.
* **× a number** — how many variations you get. Left alone, what you said wins:
  *"a new one"* gives you one, *"6 options"* gives you six.
* **Brand** — on, the run uses your workspace's brand elements: its visual
  direction and its logo. Off, it stays neutral. This one **sticks to the
  board**, so a board you set to brand-on stays that way next time you open it.
  If your workspace has no brand set up yet, the button says so and links you to
  where to add it.
* **Style** — appears once you have saved a style, and applies that look to the
  run. It rests at off, so a saved style is always a choice you make.

Size and variations are for this sitting; brand and style are decisions about
the work.

## Changing one image

Click an image on the canvas and two buttons appear on it:

* **Iterate** — change this image. Your next message is about this one picture.
* **Vary** — more like this one, without changing it.

Double-click and the image fills the room, with the conversation still beside
it. **Back to canvas** returns you.

**Point at what you want changed.** Draw on the image — a dot or a loop around
a region — and that mark goes into the message box as part of your next
request. *"Make this bit warmer"* then means something exact.

Four changes happen directly, without asking in words:

* **Crop** — trim to the part you want.
* **Re-frame** — rebuild the picture at a different shape, filling in what was
  outside the old frame. This one uses credits, and the button says so.
* **Remove background** — take the subject out of its background.
* **Vectorize** — turn it into clean line artwork that scales without going
  soft.

**Text** places a headline or a call to action on the image. It sits as a layer
over the picture rather than being burned into it, so you can move it, change
it, or take it off — and it is baked in when you export.

Every change makes a new image on the board rather than overwriting the one you
started from, so you can always go back to an earlier version.

## Finishing an image

Everything on a board is either **Draft** or **Final**, and the switch at the
top of the board shows you one or the other.

* **Mark final** says this one is done. A final image joins the Images roll in
  **Studio** on the [Boards page](/work/boards), and being final is what lets it
  be used elsewhere.
* **Save style** keeps the look of this image as a named style, which then
  appears as the **Style** button on your boards. If the image already is one,
  the button says so.

In Studio you can also **pin** an image with the star on its tile, which floats
it to the front of its roll. Pinning is a favourite and nothing more — it does
not make an image final.

## Getting the file out

**Download** saves the image as it is.

Beside it, **Export format set** saves it at every shape you are likely to
need, in one zip:

| Shape              | Where it fits                                    |
| ------------------ | ------------------------------------------------ |
| 16:9 landscape     | YouTube thumbnail · video frame · landscape hero |
| 1:1 square         | Square social · avatar · tile                    |
| 9:16 vertical      | Story · reel · short                             |
| 4:3 landscape feed | Feed landscape                                   |
| 3:4 portrait feed  | Feed portrait                                    |

## Where images go next

An image made here can be attached to a piece on [the
schedule](/work/schedule), used in a video's scene plan, or simply downloaded.
Boards hold the material; the schedule is where something becomes a thing that
goes out on a day.