Today, I tried out the Scenario platform, whose official slogan is "AI for game assets."
The website is https://www.scenario.gg/.
Currently, the main features include: Generator, Image, and Canvas.
The Generator function is similar to the model of Stable Diffusion. You just need to upload 5-100 images to train a model. I tested it by uploading 5 images for training, which took about half an hour.
The Image function uses models previously trained by users or from the public library to generate images.
The platform has already trained some models based on game scenarios.
The Canvas function is like a drawing board, providing some basic features such as uploading images, generating images, moving, wiping, and doodling. It allows planners to generate sketches through simple drag-and-drop operations. However, this feature is currently in the testing phase, and its practicality still needs improvement.
Overall, this product now feels more like a demonstration version. The images generated at present cannot be directly used for game production. However, considering that the operational threshold for training Lora using Stable Diffusion WebUI is relatively high, Scenario provides a relatively simple starting point for people to try.