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

# Bulk actions

> Process multiple repetitive tasks with a single request.

Bulk actions is a feature that applies the same operation to multiple images or files at once. You can handle simple, repetitive generation or editing tasks with a single command.

## Method 1: Apply to files generated during conversation

If you just created multiple results (e.g., 10 images) while chatting with Carat AI, you don't need to re-upload the files. Just continue requesting in the chat.

```text theme={null}
Change all 10 images I just made to 16:9 ratio.
```

```text theme={null}
Create videos from all these images with 'natural movement' content.
```

## Method 2: Upload new files and apply

Upload multiple source files that need processing at once, then request bulk actions via text.

<Steps>
  <Step title="Upload files">
    Upload 10 images you want to edit at once.
  </Step>

  <Step title="Request bulk action">
    Request the bulk action via text in the chat.

    ```text theme={null}
        Remove the background from all 10 uploaded images.
    ```

    ```text theme={null}
        Convert all these images to 'Disney style'.
    ```
  </Step>
</Steps>

<Tip>
  **Specifying results for revision**

  When multiple results (1-10) are generated through bulk actions, you can specify particular results for revision.

  * "I don't like video #3 from what you just made. Redo it."
  * "Redo only images #5 and #7."
</Tip>

## Generating multiple different images at once

Unlike bulk actions that repeat the same operation, you can also generate **multiple different images** in a single request. Just describe every scene you want in the chat.

```text theme={null}
Create 10 different images:
1. A cat surfing on a beach at sunset
2. A futuristic city at night with neon lights
3. A climber at the summit of a snow-covered mountain
...
```

After the results are generated, you can redo specific ones — for example, "Redo only #5 and #7."

<Info>
  Applying a different prompt per image via an external file such as CSV is not supported. List the different prompts together in the chat as shown above.
</Info>