Answer me with HTML

Answer me with HTML is an open-source agent skill that turns a Markdown draft into a readable, self-contained HTML page. It is designed for questions where a diagram, table, timeline, flow, code sketch, or structured explanation is easier to understand than a long block of prose. Ask your coding agent as usual; when a page is useful, the agent writes the content in a short draft and the included command-line renderer handles the layout and presentation.

The project works with Claude Code, Codex, Cursor, OpenCode, Pi, and DeepSeek Harness. The model focuses on the explanation instead of manually writing repetitive HTML, CSS, and SVG layout. The renderer can create diagrams and other structured components from a compact Markdown format, and the result is a single HTML file that can be read offline. The website describes the output as using about one eighth of the tokens that a model would need to write the same HTML by hand. Actual savings depend on the prompt, model, and setup.

Use it for technical explainers, architecture and call-chain maps, sequence diagrams, comparisons, timelines, annotated writing reviews, project plans, or other multi-part answers. A one-line question can still receive a one-line answer; the page format is intended for explanations that benefit from visual structure. The skill does not replace the model or require a separate hosted rendering service. The CLI ships inside the skill, so there is no separate npm install step. The demo includes a browser playground where a draft is rendered as you edit it.

The project is MIT-licensed and available on GitHub. Installation instructions are provided on the project site for supported agents, and Node.js 20 or newer is listed as a requirement. Generated pages are intended to be portable: one HTML file, with the layout and diagrams rendered locally. This keeps the output easy to save, share, and open later without depending on the original website.

Agent skill that turns Claude's answers into readable one-page HTML explainers. Answer me with HTML keeps the model's writing focused on the content and lets its renderer turn the draft into a compact, visual page.