url-to-markdown

Curated toolkit for turning a URL into clean markdown. None of the tools ship with this repo — install on demand.

Decision tree

Situation	Tool	Why
Default — any news article	Jina Reader (hosted)	Best markdown output, handles metadata (title, author, date), zero install.
Offline / self-hosted / no third-party calls	Mozilla Readability (via `go-readability` single binary, or `readability-lxml` Python CLI)	Battle-tested Firefox Reader View algorithm; single dependency; no network beyond fetching the URL itself.
Readability mis-extracts (list pages, unusual layouts, non-English)	trafilatura (Python)	Different algorithm, often wins where Readability loses. Markdown output built in.
Source is a PDF / .docx / .pptx (government report, ISW white paper)	markitdown (Microsoft, Python)	Handles non-HTML primary documents the other tools can't.
Page is JS-rendered and Readability returns empty	imperium-crawl (from the `rss-discover` toolkit) to fetch rendered HTML, then pipe into Readability or trafilatura	Two-step: render, then extract.