Features
Image Generation
You can ask any being or companion to generate images directly in the conversation. Switch to an image-capable model and ask naturally — no special commands.
How it works
Image generation uses vision-capable models from OpenRouter. When you ask for an image, the being generates a response that includes it inline. The image appears directly in the conversation — you can save, share, or reference it later from your gallery.
You don't need a special command. Just ask naturally, and the being will generate one if the request is clear.
Enabling image generation
Image generation requires an image-capable model. To switch:
Tap the gear icon or open Settings from the sidebar.
Find the model selector in the API / Model section.
Look for models labeled with image support — e.g., Gemini Flash or Gemini Pro with image capability. The image button in the input bar will become visible once selected.
Once an image model is active, a small image icon appears in the chat input bar. You can tap it to explicitly trigger generation, or just ask naturally.
How to ask
There are no special commands. The being detects image intent from natural language. Phrases that reliably trigger generation:
You can also describe a specific visual style, mood, or composition alongside your request — the being passes this context to the model.
Aspect ratio
Set a preferred aspect ratio in Settings → Image Aspect Ratio. Options include 1:1 (square), 16:9 (landscape), 9:16 (portrait), and more. The setting applies to all image generation until you change it.
Image gallery
Every generated image is automatically added to your gallery — a searchable, filterable archive of everything created across all your conversations.
Access it at Settings → Gallery on mobile. The gallery shows a three-column grid with filter chips for each being and companion. Tap any image for a full-screen view with pinch-to-zoom. Save or share individual images, or download all visible images at once.
Images generated using cloud models are hosted remotely and may expire after approximately 24 hours. Save anything you want to keep before then.
Using a different model per companion
If you want one companion to always use an image model and another to use a text-only model, you can set per-companion model overrides.
In the Summoning Portal, tap on a companion's settings and look for Model Override. Set a specific model for that companion — it will use this model in every conversation regardless of your global model setting.