Editing and describing images.

Every edit is a new file

When you edit an image, crop it or turn it into a format, a new file is created. The original stays exactly as it was — on every page that already shows it, too. The new file sits in the original’s stack and carries the “edited” label.

It takes over the original’s image description and source, and for free images also the licence record, with a note that it is an edit. Like every new file, it is checked on arrival.

Every version counts fully towards your storage. That is why the “Edit” bar and “Formats” show how much is “Remaining:” before you click. Only images can be edited — not PDF documents and not videos.

The image description

The image description is a short sentence saying what can be seen in the image. People who have pages read aloud hear it in place of the image, and search engines read it too. You write it once here; when you use the image in a tool, the tool takes it along.

The tile’s details show “+ image description” — or the description that is already there. A click turns it into an input field. It is saved when you leave the field or press Enter; Escape discards the change.

A good sentence names what matters: “Flat white with a milk-foam leaf on a wooden table” rather than “coffee”. Which images still lack a description is shown under “Tidy up” in the “No description” view.

1. 1Open the tile’s “Details”.
2. 2Click “+ image description” and write what can be seen.
3. 3Press Enter or click elsewhere — the description is saved.

Letting Kenji suggest a description

If no sentence comes to mind, let Kenji suggest one: in the details, click “Let Kenji suggest a description”. Kenji looks at the image and writes a sentence in the language of your interface. It then stands in the input field — not saved yet.

Read the suggestion, change whatever is not right and leave the field: only then does it become your image description. Escape discards it. Kenji is instructed not to guess the names of people, places or brands and to mention only text that can really be read in the image.

A suggestion uses AI credit; saving doesn’t. If the suggestion fails, nothing is charged. If your AI bar is empty, the message tells you: it refills on its own, or you add an AI pack (see Account & billing).

1. 1Open the tile’s “Details” and click “Let Kenji suggest a description”.
2. 2Read the sentence in the field and change it where needed.
3. 3Leave the field or press Enter — now the description is saved.

Editing in the editor

The editor opens an image in a window over the library. It has five areas: “Crop & rotate”, “Fine-tuning”, “Filters”, “Annotate” and “Resize”. The editing happens in your browser — only the result is uploaded.

You open it with the “Edit” pencil on the tile, with “Edit” in the large window, or in the details via “Edit” → “Open editor”. A version can be edited further the same way.

“Save” opens a window where you set the name and file format of the new file; “Save” in there creates it. The editor stays open until the file really exists — if saving fails, for instance because your storage is full, your work is not lost. “Close” leaves the editor without a new file.

1. 1Click the “Edit” pencil on the tile.
2. 2Choose the section you want under “Crop & rotate”.
3. 3Click “Save”, check the name and format, and confirm with “Save”.
4. 4The new file then sits in the original’s stack.

Removing backgrounds and enlarging

In the details, “Edit” opens a bar. Next to “Open editor” it offers two actions that need no editor: “Remove background” takes the background away and trims the empty border along with it — handy for product photos. “Enlarge” brings an image up to the next size and generates new pixels instead of stretching the existing ones.

Both create a new file in the stack; an image with its background removed becomes a PNG with a transparent background. If an image is already as large as the largest size, the message says so instead of creating an identical file. If an action does not work, no file is created, and the message gives the reason.

Removing backgrounds and enlarging cost no AI credit, but they do take up storage.

1. 1Open the tile’s “Details” and click “Edit”.
2. 2Click “Remove background” or “Enlarge”.
3. 3After “Processing …”, the new version appears in the open stack.

One image, every format

Every tool needs the same image in a different shape: the website a wide header, a post a portrait shape, a story a tall, narrow one. “Formats” in the details creates up to five crops from one image at once: “Story and Reel”, “Portrait post”, “Square post”, “Website header image” and “Share preview”. Dimensions and aspect ratio are shown next to each; “Story and Reel”, “Square post” and “Website header image” are preselected.

The crop is made around what matters in the image, not rigidly in the middle. Each crop becomes a file of its own in the stack and counts towards your storage; a format that already exists is not created twice. Formats cost no AI credit.

The crops are then available wherever you place an image — in the tools’ image picker, among the versions of their original.

1. 1Open the tile’s “Details” and click “Formats”.
2. 2Tick the formats you need.
3. 3Click “Create … formats”, such as “Create 3 formats”. The crops appear in the stack.