wtf is 3d-prompt-collection?

3d-prompt-collection is a curated list of text prompts meant to be copied and pasted into an AI coding tool to generate ambitious three dimensional scenes, mostly built with the Three.js JavaScript library. The README itself is really just the prompt library: each entry is a title that expands to reveal a long, detailed paragraph describing exactly what the generated 3D world should look like, down to camera movement, lighting, and how the scene should hold together as one continuous zoomable object rather than separate disconnected pieces. The collection is organized into six themed sections. Big 3D Worlds is the largest, with thirty prompts describing entire recreated cities and historic sites, such as all of Manhattan Island, ancient Rome, the Seven Wonders on one horizon, and the Forbidden City in snow. Playable and Game-Like Scenes covers a dozen prompts for more interactive feeling settings like rooftop parkour across New York or a flight simulator over the Alps. Living Art Worlds turns famous paintings, including Van Gogh's Starry Night and Monet's water lilies, into explorable three dimensional scenes. Impossible Vantages, Natural Spectacles, and an Elemental and Cosmic Finale section round out the collection with surreal perspectives, real natural events like Niagara Falls and monarch butterfly migrations, and large scale elemental scenes such as a volcanic island being born. Every prompt is written to push a coding assistant toward extreme scale and detail, repeatedly telling it not to fake depth with a flat skybox or a handful of buildings, and instead to use techniques like instancing and level of detail so that thousands of individually placed objects can render smoothly. There is no application code to install and no dependencies to configure. The repository is simply a long, single markdown file of copyable prompt text, intended for anyone experimenting with AI-generated 3D scenes to reuse directly. The full README is longer than what was shown.