ONE MCP CONNECTION · IMAGE
AI image generation & editing with MCP
Create campaign visuals, illustrations and product concepts from a prompt. Edit an existing image or use references, without managing your own image API key.
What you can do
- Generate an image from a text description.
- Edit an image using a prompt and reference images.
- Request a transparent background with a model that supports it.
- Choose your model in your account; output size and quality depend on the provider.
Try this brief
Example prompt · not a claim that this output has already been generated
Create a square editorial illustration of a quiet coffee shop at sunrise. Use warm cream and forest green, with space for a headline. No text.
Available tools: image_generate / image_edit
Models and pricing
| Model | Starting rate |
|---|---|
| GPT Image 2.5 Sunburstgpt-image-2.5-sunburst | 6 credits / image |
| GPT Image 2.5 Flaregpt-image-2.5-flare | 6 credits / image |
1 credit ≈ US$0.01. Settings and reference inputs can affect the final quote. Current account pricing takes precedence.
Before you start
The current image provider can treat size and quality as hints. The service reports the actual file format; a requested setting is not a guarantee of an exact result.
Connect with remote MCP over Streamable HTTP and OAuth. Google sign-in creates your Lavis account. Interactive previews require MCP Apps support; download files you want to keep before their 7-day expiry.