Skip to main content
Image generation turns a text description into an image URL or file you can send back in the chat. Some providers also support image editing.

​
Providers

Connect keys under AI Providers.

​
Generate an image

1

Collect the prompt

Use User Input (or a static prompt) and save the description to a custom field.
2

Add the provider action

Choose Generate image for OpenAI, Gemini, or xAI.
3

Map input and output

Pass the prompt field in; save the result URL/file to another custom field.
4

Send the image

Use an Image / Message block that references the output field.

​
Edit an image (Gemini)

  1. Collect the source image into a custom field.
  2. Collect edit instructions (“remove background”, “add a red circle”).
  3. Run the Gemini edit image action with both fields.
  4. Send the edited output.

​
Quality tips

  • Be specific about subject, style, aspect, and what to avoid
  • Remind the model about brand constraints in the prompt
  • Verify channel media limits before sending large outputs
  • Keep a fallback message if generation fails
Generated images may be inappropriate or off-brand. Add moderation/review steps for customer-facing creative workflows.

​
Next

Vision

Analyze images customers upload.

Rich responses

Send carousels and buttons around generated media.