Skip to main content

​
Enable image generation

Open your Verb’s Image Engine, enable generation, and choose a model. The selector shows available models and their costs. Your plan’s image allowance and any model credit charge are separate; see Plans and credits.

​
Ask for an image

In a conversation, try:
Use a clear request to create an image, not just a description of a scene. In connected platforms you can also use:
  • Discord: /generate with a prompt, optional ratio, and optional reference attachment.
  • Telegram or Slack: /generate <prompt>.
  • WhatsApp: /generate <prompt> or a natural-language request.
  • Twitch: !generate <prompt>.
  • Email and Verba Chat: ask in plain English.
Verba enhances the prompt into a more descriptive brief and gives the image an AI-generated title. The final prompt is visible in Gallery.

​
Use reference images

Attach an image and say what to keep or change: “Keep this person’s appearance, but place them in a snowy forest.” In Telegram, you can reply to a photo with /generate. Where a platform doesn’t offer attachments, include an accessible image URL in the request. Twitch uses links. The API accepts image_urls. Use references you have permission to share; private or expired links may not be readable. When you explicitly ask to show your Verb itself, its avatar and banner can help guide its appearance. Results won’t always reproduce a reference exactly.

​
Choose an aspect ratio

Say square, landscape, or portrait, or request a ratio such as 1:1, 16:9, or 9:16. The selected model may adapt the exact size to what it supports. Twitch returns an image link; other supported messaging integrations send an attachment or inline image. New generated images are also saved in Gallery.
The public image API currently supports square 1024x1024 output only. Its options are narrower than interactive chat.
If generation fails, check that Image Engine is enabled, the model is available to your plan, and you have allowance and credits remaining. See Troubleshooting.