Zoo

Zoo is an open-source experimentation platform for AI image generation, designed for developers and curious users to test machine learning models in a visual interface. Built on Replicate's infrastructure, it removes the need for local GPU setup while providing access to cutting-edge models like Stable Diffusion and ControlNet. Users can input text prompts, adjust technical parameters like sampler methods and guidance scales, then compare results side-by-side using the X/Y plot feature. The tool particularly benefits AI researchers exploring model behaviors, developers prototyping image generation features, and tech-savvy creators wanting more control over their AI art outputs. While powerful for experimentation, it requires some understanding of ML parameters and relies on Replicate's backend for model execution.