What is this?
It's an Antigravity (AGY) agent skill that plans, designs, and builds video-driven parallax hero landing pages from a single prompt, using Gemini's video and image generation models.
It's an Antigravity (AGY) agent skill that plans, designs, and builds video-driven parallax hero landing pages from a single prompt, using Gemini's video and image generation models.Is this a standalone app or something I add to an agent?It's a skill definition made for the Antigravity agent; it can also run in other agent environments with minor adaptations, but it isn't a standalone application.
It's a skill definition made for the Antigravity agent; it can also run in other agent environments with minor adaptations, but it isn't a standalone application.How do I install and use this skill?Install it globally or into your current workspace, or skip installation entirely by having your agent read the repo directly from the chat URL; then invoke it with its slash command or a natural-language prompt describing the page you want.
Install it globally or into your current workspace, or skip installation entirely by having your agent read the repo directly from the chat URL; then invoke it with its slash command or a natural-language prompt describing the page you want.Do I need a Gemini API key?Yes, a Gemini API key is required for the Omni video generation step, obtainable from Google AI Studio.
Yes, a Gemini API key is required for the Omni video generation step, obtainable from Google AI Studio.Does generating a page cost anything beyond the skill itself?Yes, each build calls paid Gemini models (gemini-omni-flash-preview for 10s video, Nano Banana for reference images), so using it consumes your own Gemini API credits.
Yes, each build calls paid Gemini models (gemini-omni-flash-preview for 10s video, Nano Banana for reference images), so using it consumes your own Gemini API credits.How does the skill actually build the landing page?It generates reference images with Nano Banana (gemini-3.1-flash-image), creates a 10-second 16:9 video with gemini-omni-flash-preview, assembles a 400vh sticky video-scrubbing index.html with progress-locked typography, then launches a local preview server.
It generates reference images with Nano Banana (gemini-3.1-flash-image), creates a 10-second 16:9 video with gemini-omni-flash-preview, assembles a 400vh sticky video-scrubbing index.html with progress-locked typography, then launches a local preview server.Is this an official Google product?No, the README explicitly states it is not an official Google product, even though it uses Gemini and Nano Banana models.
No, the README explicitly states it is not an official Google product, even though it uses Gemini and Nano Banana models.What kind of design does it avoid?It enforces clean design rules by skipping unnecessary icons, fake tickers, and toy widgets so the page stays minimal and content-first.
It enforces clean design rules by skipping unnecessary icons, fake tickers, and toy widgets so the page stays minimal and content-first.Can I use it outside Antigravity?Yes, it can run in other agent environments with minor adaptations, though it's built and tested primarily for Antigravity.
Yes, it can run in other agent environments with minor adaptations, though it's built and tested primarily for Antigravity.What are similar projects?kid-papercraft, gemini-skills, taste-skill, unreel, handraw-style, dioramas, cookbook, generative-ai
kid-papercraft, gemini-skills, taste-skill, unreel, handraw-style, dioramas, cookbook, generative-aiWhich topics is it in?nano-banana, gemini-api, landing-page, web-design, prompt, antigravity, google, video-generation, image-generation, design, skill, video
nano-banana, gemini-api, landing-page, web-design, prompt, antigravity, google, video-generation, image-generation, design, skill, video