User can click "Check archive.ph manually →" from inside Miniflux to see if a snapshot exists or to trigger one. Useful when the automated archive.ph lookup hit rate limits or no snapshot was ready yet.
- crawl4ai direct first (detect if article actually needs paywall bypass)
- Only hit archive.ph when paywall detected (saves rate-limit budget)
- New 📰 icon for free articles enriched with full text
- 📖 reserved for actual paywall bypass via archive.ph
Final state set:
- 🔓 OA via Unpaywall
- 📰 Free article enriched
- 📖 Paywall bypassed via archive.ph
- ⏳ Paywalled, retrying
- 🚫 Paywalled, gave up after 3h
- ❌ Fetch error
Replace 6-min published_at cutoff with emoji-suffix guard so the
scheduled flow processes all unread paywalled articles regardless of age.
Fixes badge text: "Wallabag" -> "crawl4ai".
Remove Wallabag token node; fetch full text directly via crawl4ai /md.
Scientific articles DOI-resolved via Unpaywall API for open-access URLs.
Adds title emoji suffix: 🔓 (OA), 📖 (crawl4ai), ❌ (failed).
Includes backfill variant (no time cutoff) for processing existing unread entries.
Runs every 5 min. Detects paywalled articles (known domains + short content).
For scientific: tries Unpaywall API (DOI-based OA lookup) first, falls back to
Wallabag. Writes full text back to Miniflux entry so reader shows it inline.
fkrebs