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

# Media viewer

> Open any generation full screen, zoom, inspect the request, mark it up, and send an edit straight back to the agent.

Click a tile to open the media viewer. It shows one generation at full size with a filmstrip of the others in the chat. The header shows the ordinal, the name, the format, and the dimensions.

<Frame>
  <img src="https://mintcdn.com/fal-d8505a2e/DmEWlE6WV9vaz0Me/images/agent/media-viewer.jpg?fit=max&auto=format&n=DmEWlE6WV9vaz0Me&q=85&s=5010b19e775ef8318c35d040d9c1f8dc" alt="The media viewer with the filmstrip, the image toolbar on the left, and the edit composer at the bottom" width="1456" height="840" data-path="images/agent/media-viewer.jpg" />
</Frame>

## Controls

| Control           | Action                                                                                                                      |
| :---------------- | :-------------------------------------------------------------------------------------------------------------------------- |
| Previous and next | Move through the filmstrip                                                                                                  |
| Zoom              | Zoom in and out on images. The reset control returns to 100%                                                                |
| Favorite          | Toggle the heart                                                                                                            |
| Request details   | Open a panel with the endpoint, request ID, status, attempt, and duration. Use **Copy request id** when you contact support |
| Close             | Return to the chat. `Escape` also closes                                                                                    |

## Edit from the viewer

The composer at the bottom of the viewer sends an instruction to the agent with the current image attached. Type an edit, or ask for a variation, without leaving the viewer.

The left toolbar offers quick actions for image outputs.

* **Edit** opens the markup canvas. See below.
* **Quick animate** turns the still into a short video.
* **Upscale** runs an upscaler on the image.
* **Make a variation** asks for one or more new takes. A picker sets the count.

## Markup and crop

**Edit** opens a drawing canvas over the image. Use it to point at the region you want changed.

<Frame>
  <img src="https://mintcdn.com/fal-d8505a2e/DmEWlE6WV9vaz0Me/images/agent/markup-canvas.jpg?fit=max&auto=format&n=DmEWlE6WV9vaz0Me&q=85&s=4a7adb71a4869e98484fe140375a5313" alt="The markup canvas with pen, highlighter, eraser, arrow, text, and crop tools, and a red stroke drawn on the image" width="896" height="1344" data-path="images/agent/markup-canvas.jpg" />
</Frame>

| Tool                | Use                                             |
| :------------------ | :---------------------------------------------- |
| Pen and highlighter | Draw on the image. Pick a color from the swatch |
| Eraser              | Remove a stroke                                 |
| Arrow               | Point at a region                               |
| Text                | Label a region                                  |
| Crop                | Keep a region of the image                      |
| Undo, redo, clear   | Step back or start over                         |

Click **Done**. The canvas is flattened to a PNG and placed in your composer as an attachment. Type what you want changed and send. The agent uses the marked image as the edit source. **Cancel** discards the marks.