AI in practice

Stableboost generates up to 200 images in one go with Stable Diffusion

Jonathan Kemper

Screenshot

Even with a perfect prompt, there are practically an infinite number of ways an AI-generated image can turn out. The website Stableboost shows up to 200 of them after just one run.

DALL-E 2 and Midjourney generate four images per run, Dreamstudio, StabilityAI's official web app for Stable Diffusion, even up to nine - but even then, it can take quite a while to generate the right image.

Just two months after Stable Diffusion's release, the market for browser-based Stable Diffusion services has grown rapidly. The website stableboost.ai is fundamentally different from its competitors: Stableboost is designed from the ground up to allow users to quickly generate many images, says developer Tal Stramer. The number of images can be set between one and 200.

Direct multiple variations with modifiers

Apart from functions such as image-to-image, inpainting, face restoration and upscaling, which are now part of the standard repertoire of such services, the so-called modifiers are a special feature of Stableboost. This allows you to specify different terms with which the prompt should be combined right from the start.

According to the developers, Stableboost allows you to interactively discover your favorites after generation. Once you have selected a few favorites, the system recommends similar images.

In addition to single images, Stableboost also supports rudimentary motion pictures. With the video function, you can define up to five prompts, which Stableboost dynamically stitches together into an animated image.

You get 500 free credits per month

Stableboost requires registration, but you can try it for free with 500 inclusive credits per month. Developer Tal Stramer wants to offer the service completely free of charge. To cover his costs, he charges one cent for each additional credit. Those who can't afford that should contact him for more free credits.

Sources: