readdown
Fetch web pages and convert HTML to clean, LLM-optimized Markdown. Extracts article content, strips boilerplate, estimates token count. No b
Open source Open in the app JSON README (API)
About
Fetch web pages and convert HTML to clean, LLM-optimized Markdown. Extracts article content, strips boilerplate, estimates token count. No browser needed.
Details
- Kind
- Plugins
- Topic
- AI, RAG & memory
- Publisher
- zcag
- Origin
- marketplace
- Category
- ferramentas
- Last push
- 2026-03-12T06:12:51Z
- Repository state
- ativo
- Added
- 2026-08-30 01:48:58
- Updated
- 2026-08-30 01:48:58
- Origin id
zcag/readdown-plugin/readdown
README
# readdown — Claude Code Plugin Fetch web pages and convert HTML to clean, LLM-optimized Markdown directly in Claude Code. ## What it does - **Fetch any URL** and get clean Markdown with article content extracted - **Convert HTML** strings to Markdown - **Token estimation** — know how much context window a page will use - **Metadata extraction** — title, author, date - Strips navigation, ads, sidebars, and boilerplate automatically ## Install In Claude Code, run: ``` /install-plugin https://github.com/zcag/readdown-plugin ``` Or add manually to your Claude Code settings. ## Usage Once installed, Claude can fetch and convert web pages: > "Fetch https://example.com and summarize it" > "Convert this HTML to markdown: <html>..." > "How many tokens is this article?" The plugin provides two MCP tools: - **fetch_markdown** — fetch a URL and convert to Markdown - **convert_html** — convert an HTML string to Markdown ## How it works Uses [readdown](https://github.com/zcag/readdown) — a single-library replacement for `@mozilla/readability` + `turndown`. Extracts article content and converts to clean Markdown optimized for LLMs. - No browser needed (uses linkedom, not Playwright/Puppeteer) - Fast (~3ms per article) - Small footprint (5KB core vs 65KB for readability+turndown) ## License MIT