Reverse lookup XKCD comics using Gemini multimodal embeddings
Search XKCD by image or text description using Gemini multimodal embeddings.

Beautiful AI-assisted comic, but it's a portfolio piece, not a product or tool.
Comic artists, indie creators, AI enthusiasts interested in creative workflows
Search XKCD by image or text description using Gemini multimodal embeddings.
LongTextBench 0.9733 beats FLUX on text rendering, but image gen is a crowded battlefield.
Character consistency is better than raw generators, but still just another AI comic wrapper.
Drop-in reader that actually respects Japanese reading conventions: RTL page order, automatic two-page spreads in landscape, and touch-friendly features like pinch zoom with inertia and double-tap zoom. It’s deliberately tiny — ES module, no build step, localStorage resume — and adds practical niceties (insertable ad/pages, preview limits, status-bar cover) that make it useful for anyone shipping comics on the web, even if the idea itself isn’t groundbreaking.
Hand-drawn AWS comics with genuine charm and personality.
Four hours of vibe coding with auto-commits as safety nets.