Best open-source HTML projects

A collection of 55 installable skills that teach AI coding agents to build motion graphics, animation and video, organized into 18 packs.

A set of installable Claude Code and AI agent skills covering motion design principles, After Effects, Remotion, logo animation, and motion backgrounds.

A free, self-paced path through a full computer science curriculum built from online courses and books, with community support.

A self-hosted web app for downloading videos and audio from 1000+ sites via yt-dlp, exporting as MP4 or MP3.

A curated survey and paper list covering AI-assisted automated scientific research across the full lifecycle, from idea generation to dissemination.

A community playbook for running autonomous AI coding loops with Geoff Huntley's Ralph methodology, covering specs, planning, and build prompts.

An open-source, login-free catalog of interactive Three.js/WebGL UI components with live renderers and full source code.

CLI and MCP server that extracts a website's full design system (colors, type, spacing, motion) from the live DOM into tokens and framework-ready files.

A pure front-end floor plan tool for arranging furniture, editing walls and switching to a Three.js 3D walkthrough, runs from a single index.html.

An offline-first, MIT-licensed learning platform bundling competition problems, exam prep, civil service materials, Mao Zedong texts and WorldSkills courses in one app or website.

A free, open-source Chinese guide for people who just lost their job, covering severance, unemployment benefits, medical insurance, housing fund, job search and cash flow.

A single-page static site showing a Chinese character's oracle bone, small seal, traditional and simplified forms with Shuowen, pronunciation and phonetic-family data.

Agent skills that guide coding agents to build self-contained HTML wireframes, interactive prototypes, plans, and diagrams.

AI-powered news digest that turns Twitter, RSS and other feeds into structured summaries on a web dashboard.

A Python benchmark that measures how well LLMs solve the Bulls and Cows number-guessing game, using LiteLLM to reach many model providers.

A domain-general framework that uses world-model predictions and interaction feedback for recursive self-improvement across code repair, GUI tasks and lab planning.