Web Image Downloader

Problem

When images are scattered across one or more pages, links, ads, duplicates, and low-quality files make manual collection hard. This skill turns URLs into a controlled capture workflow for asset collection, backups, reference libraries, and logged-in page images.

How It Works and Limits

It processes the page in stages: fetch HTML, parse image links, detect galleries, remove duplicates with MD5 and perceptual hashing, filter ads, and drop low-quality images using quality thresholds for blur, low color richness, or large watermarks. It supports Cookie strings, Netscape Cookie files, resume, concurrency, size filters, and can narrow results by minimum and maximum pixel dimensions to reduce unrelated small images or oversized originals. It also accepts comma-separated URLs or URL files. Output is organized into format folders, AI category folders, report.json, session.json, and a ZIP archive; the report includes quality scores and category labels. It is intended for public or already authenticated web resources and does not guarantee bypassing strong anti-bot controls, signed URLs, or paywalls; AI categorization is heuristic and best used as a starting point for manual review.