A breakdown of the duplicate finder and its 15+ perceptual signals, from dHash, aHash, and pHash to texture, color, and shape measures that catch visual copies.
How the duplicate finder sensitivity presets (Strict, Balanced, Aggressive) change how aggressively visual duplicates are grouped, and which one fits your batch.
A practical guide to tune duplicate finder weights: adjust per-signal importance, re-apply without rescanning, and fix matches that are too loose or too tight.
How the 404 checker validates image URLs with lightweight HEAD requests, sorts reachable from unreachable, and lets you export a clean list to re-import.
How to find the best parallel download setting: start at 5-10, raise concurrency gradually, pair high counts with ZIP mode, and reduce if a server blocks you.