01 // DISCOVERY CACHE
Build the manga catalogue once. Search it instantly.
The automatic ‣ loop fetches successive batches, upserts records into IndexedDB, updates existing cache entries, and stops on a terminal batch or repeated no-new results.
Parallel request count and launch gap are controlled from Settings. Cache exports are portable password-protected .arjin archives; legacy JSON imports remain supported.
Ready • cache duration follows Settings.
Manga metadata
0
Chapters
0
Page bytes
0
02
Advanced manga search
0 cached mangaLoading cache…
03
Loop control
0 BRemote .arjin sync uses the password and source URL in
config/site-links.json. GitHub blob URLs and standard Google Drive file URLs are normalized automatically; other providers must expose a direct-download URL.Normal users can run the ‣ control without touching technical settings. The homepage now stores metadata in independent alphabetical cache shards so first paint does not require a 10,000+ record catalogue read. The loop resumes from the saved next offset, launches multiple batches in parallel, waits only the configured launch gap between request starts, and merges each window together. Imported resume-safe JSON restores its cursor; older JSON without a cursor uses a safe count-based fallback. Source mapping, bridge/CORS policy, parallel batch count and launch gap live in Settings.
Advanced loop controls
These controls mirror the Discovery source adapter. Save source changes from Settings.
LOG