Ostaran Scrape — web reader (no AI)

Path style (prefix any URL):
  GET /https://example.com/article
  Headers: X-Format: md|txt|html|json, X-Render: 1, X-Refresh: 1, X-Download: 1

Query style:
  GET /read?url=https://example.com&format=md&render=1&download=1

Full-site crawl (downloads a ZIP of markdown):
  GET /crawl?url=https://example.com&maxPages=50&maxDepth=3&format=md
  Discovers same-domain pages via sitemap.xml + link-following.

Formats: md (default), txt, html, json
render=1   -> use headless Chromium for JS-heavy pages (requires Playwright)
refresh=1  -> bypass cache
download=1 -> send as a file attachment