Logo of the AI Catalog website
Logo of the AI tool "Say What You See"

Say What You See

AI Image Prompting Game

Say What You See is an interactive AI experiment that trains users in the art of image prompting by challenging them to describe generated images accurately. Players receive feedback on their semantic matches to improve visual literacy and prompt engineering skills.

Visit

More AI Tools for Fun & Experiments

View All
Logo of the Xinghui (星绘) AI tool

AI Avatar and Style Generator

Xinghui (星绘) is an AI image generation platform that allows users to create personalized AI avatars and transform photos into various styles like cyberpunk or manga. It enables easy customization of visual identities through simple keywords and image uploads.

Logo of the Learn Everything AI tool

Visual Analogy Learning Tool

Learn Everything is an AI experiment that generates personalized explanations of complex topics using your own photos as visual metaphors. By analyzing objects in your surroundings, it creates relatable analogies to explain subjects ranging from quantum mechanics to biochemistry in a style that suits you.

Logo of the Talking Tours AI tool

Generative Audio Tour Guide

Talking Tours is an AI experiment from Google Arts & Culture that generates real-time audio guides for cultural sites explored in Street View. It uses dynamic visual analysis to provide spoken insights and answer user questions about landmarks.

Logo of the Opal AI tool

No-Code AI App Builder

Opal empowers anyone to discover, build, and deploy AI mini-apps using a visual editor and natural language prompts. This no-code platform turns creative ideas into working applications in minutes without looking at a single line of code.

Discover Other Tools

View All
Logo of the Whisk AI tool

AI Image Blending Tool

Whisk is an AI creative tool that enables users to blend visual inputs—subject, scene, and style—into new images using a drag-and-drop workflow. It leverages Gemini to analyze inputs and Imagen 3 to generate high-quality, concept-blending results without relying solely on text prompts.