Image Generation
The Image Generation ability lets the agent create pictures from a text description, for example a simple diagram or an illustration for a handout. Your workspace must have the AI image generation feature turned on.
Settings
| Setting | Description |
|---|---|
| Default aspect ratio | The shape the agent uses when the person does not ask for one: Auto (agent decides), Square (1:1), Landscape (16:9), Vertical (9:16), Standard (4:3) or Portrait (3:4) |
| Maximum images per request | 1 to 4 images for each request (4 by default) |
How It Works
Each generated image is saved as a private File in your workspace. The Member the agent is acting for needs permission to write Files; without it, the agent explains that it cannot create images.
After it creates an image, the agent shows it in the Chat:
- Web chat shows the image inline.
- SMS, WhatsApp and email can deliver an image only when its File has link sharing on and access set to Public. A new generated image is private, so the agent cannot send it on these channels until someone changes the File's sharing. Anyone with the link can open a Public File, so do this only for images that hold no patient information. See How SMS works.
- Phone calls cannot show images; the image is still saved as a File.
Images are made by a third-party model (Google Gemini). Keep patient details out of image requests.
Adding Image Generation
- Open your workflow or task
- Go to the Abilities tab
- Click Add Ability and select Image Generation
- Choose the settings above
- Save