# Extract Page Copy > Use when evaluating, reviewing, or critiquing copy or messaging on an HTML file or URL. Also use when the user wants to read, assess, or discuss the text on a page without wading through markup. - Skill: `darrenbounds/extract-page-copy` (Agent Skill, multi-file: 2 files) - Install (CLI): `npx skillmds@latest add darrenbounds/extract-page-copy` - Raw SKILL.md: https://api.skillmd.com/api/skills/darrenbounds/extract-page-copy/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Web & Frontend - Author: darrenbounds (https://skillmd.com/u/darrenbounds) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/darrenbounds/extract-page-copy --- # Extract Page Copy Strip all markup, scripts, and styles from an HTML file and output only the text a visitor would read — grouped by semantic role. This reduces token usage by up to 98% compared to passing raw HTML. ## When to Use - User says "review the copy on this page" with an HTML path or URL - User asks "what does this page say" pointing to a file or webpage - You need to evaluate messaging before providing feedback — raw HTML wastes context ## How to Extract Read the HTML source (from a file or URL), then apply these rules in order: **Strip completely — do not include any text from:** - `