Images - Cloosphere Guide

Admin › Settings › Images

Configure AI image generation.

The tab is split into the Image Generation switch that turns the whole feature on and off, and the Image Generation API connection list where the engines that actually produce images are registered.

Images settings tab — Connect image generation engines and enable for agents

Enable Image Generation

Setting	Description	Default
Image Generation	Image generation feature ON/OFF	Disabled

Image generation is only exposed to users when enabled as an agent capability.

Image Generation API

Engines are managed as a connection list. Multiple engines can be registered at the same time.

Action	Location
Add connection	The ＋ button at the right of the list → Add Image Connection modal
Edit connection	The gear button at the right of each row → Edit Image Connection modal
Delete connection	Delete at the bottom left of the edit modal

Each row shows engine · connection name · address (for example Azure OpenAI gpt-image-1.5 • https://…). If the connection name is left blank the model ID is shown instead, and Vertex AI shows project ID (location) instead of an address.

Edit Image Connection modal — Edit Image Connection — every per-engine setting lives inside this modal

Common settings

Setting	Description	Default
Engine	OpenAI · Azure OpenAI · Gemini · Vertex AI	`OpenAI`
Connection Name	Name shown in the list (e.g., `DALL-E 3`, `GPT Image`)	-
API URL	Engine endpoint. Not shown for Vertex AI	-
API Key	Authentication key. Not shown for Vertex AI	-
Image Size	Generated image size	`1024x1024`

Per-engine settings

OpenAI
Azure OpenAI
Gemini
Vertex AI

Setting	Description	Example
Model ID	Image generation model	`dall-e-3`

The default API URL form is https://api.openai.com/v1.

Setting	Description	Options · Default
Deployment Name	Azure OpenAI deployment name	e.g., `gpt-image-1`
API Version	Azure OpenAI API version	`2025-04-01-preview`
Quality	Image quality	`auto` (default) · `low` · `medium` · `high`
Format	Output format	`PNG` (default) · `JPEG` · `WebP`
Background	Background handling	`auto` (default) · `transparent` · `opaque`

The default API URL form is https://<resource-name>.openai.azure.com. A deployment name is used instead of a model ID.

Setting	Description	Example
Model ID	Gemini image generation model	`imagen-3.0-generate-002`

The default API URL form is https://generativelanguage.googleapis.com/v1beta. Authentication uses a Google AI Studio API key.

Setting	Description	Default
Project ID	Google Cloud project ID	-
Location	Region	`us-central1`
Model ID	Image generation model	e.g., `imagen-3.0-generate-002`
Service Account Key (JSON)	Full service account key	-

Vertex AI authenticates with a service account, so the API URL and API Key fields are not shown.

Agent Integration

For image generation to work, all 3 conditions must be met:

Image generation enabled + engine connected in this tab
Image generation capability set to Default On or Default Off in agent settings
User manually toggles image generation in chat (when Default Off)

Image generation operates as an agent Tool. When the AI decides “an image is needed”, it auto-invokes the image generation tool.

​Enable Image Generation

​Image Generation API

​Common settings

​Per-engine settings

​Agent Integration

Enable Image Generation

Image Generation API

Common settings

Per-engine settings

Agent Integration