Topaz Labs has released Bloom, a new web-based image upscaler for AI and CG content
Topaz Labs has released Bloom, a new AI-powered web application for upscaling creative digital images. Designed for use with AI-generated visuals and CG artwork, Bloom supports PNG and JPG formats and offers upscaling of up to 8x, incorporating generative detail enhancement using custom diffusion models.
Unlike traditional upscalers, Bloom introduces a process the company calls “creative upscaling,” which allows the tool to add new detail and textures based on the content and optional prompts. The system uses in-house AI models developed by Topaz’s research team and builds on prior technologies such as Redefine in Gigapixel and diffusion-based recovery systems like Recover and Starlight.
Users can select among five creativity levels, from “Subtle” to “Max,” which influence how much the tool alters and enhances the original image. A recursive detail enhancement toggle is also available for higher creativity levels. Prompts can be used to guide the upscaling output, though they are optional.
Bloom operates exclusively as a web application and processes images via cloud-based GPU clusters. It supports rendering up to four image variations at once, allowing users to compare different results. Topaz Labs offers a free plan with ten image upscales per month and paid plans that provide unlimited usage. The tool is distinct from the company’s other products like Photo AI and Gigapixel, which remain recommended for photographic content and restoration.
To find out more, visit Topaz Labs’ website.
Will say I already love this, and may subscribe later today. My only concern is the commercial usage restriction of the lower lever plan. I am a hobbyist, and most of the stuff I do is for my own entertainment. But I’d like to use generative (and non) assets to maybe do some kind of game or app in the future. If I bought the commercial usage plan later would it be OK if I did used something I’d previously generated when subscribed to the other plan? I can’t really know what I might want to do in the future with images I haven’t made yet.