Back to directory

HTML Anything

A local-first, agentic HTML editor where a local coding-agent CLI writes the HTML from more than 70 skill templates across nine deliverable surfaces such as magazines, decks, posters and social cards, with sandboxed preview and one-click export.

HTML Anything

HTML Anything is an agentic HTML editor: instead of a person hand-editing a document, a local coding-agent CLI writes the HTML directly, and the editor treats HTML, not Markdown, as the final reader-facing format. It runs local-first with no API key, auto-detecting whichever of nine supported coding-agent CLIs, including Claude Code, Cursor Agent, Codex, Gemini CLI, GitHub Copilot CLI, OpenCode, Qwen Coder and Aider, is already on the user's PATH, and reuses that CLI's existing logged-in session.

It offers 75 composable skill templates across nine deliverable surfaces: magazine articles, keynote decks, resumes, posters, Xiaohongshu cards, tweet cards, web prototypes, data reports, and Hyperframes videos. Each skill ships a real example file that opens directly from the repository with no setup, and generated output can be previewed in a sandbox and exported with one click to WeChat, X or Zhihu, or downloaded as an HTML file or PNG.

It is built by the team behind a related, larger product, Open Design, and is written primarily in HTML under the Apache-2.0 license.

More in Agent skills

See all 581
  • Provenance gate

    github.com

    A deterministic authorization gateway that sits between an AI agent and its tools, tracing where each parameter of a side-effect call came from, and blocking the call or asking a human when the source does not check out.

  • A markdown framework of system rules that restates Isaac Asimov's Four Laws of Robotics as a set of prioritized loss calculations for large-language-model agents to follow, published as a discovery file for agents and crawlers to read.

  • Vitruvius

    github.com

    An engineering research agent that runs a discover, read, synthesize, verify and review loop over technical questions, reading sources directly rather than inferring from titles, and publishing its own scored verifier benchmark, weaknesses included.