# Lazy-loaded Image URL Test — 14 October 2026

Source page: https://freqdec.github.io/peekaboo/demo/srcset-lazyload/index.html
Scanner: Source Image Finder 0.5.186 .64 — Scan Single Page

1. Download the initial public HTML without executing JavaScript (HTTP response 200).
2. Parse the actual HTML DOM for <img> elements, ignoring JavaScript strings and HTML comments. The saved HTML has 0 actual img elements; a naive regex erroneously counted an example inside script/comment text.
3. Scan the same URL using Source Image Finder's Single Page workflow; the demo creates image elements with JavaScript and stores responsive candidates in data-srcset.
4. Export selected image URLs as TXT and image details as CSV. Both exports contained the same 50 URLs, in the same order. The CSV has 50 rows and no exact repeated best_verified_url values.
5. Download the 50 selected images as ZIP. Verify the ZIP CRC and each JPEG. All 50 files were present and readable, measured 1200x675, and had distinct SHA-256 hashes. The CSV reported no larger verified source files.
6. Use the manifest JSON to check the recorded filenames, file sizes, image URLs, and SHA-256 checksums. JPEGs are not redistributed by Source Image Finder.

Limitations: One public demonstration, one captured run, and a deliberately simple non-rendering HTML baseline. The results do not establish universal lazy-loading support, performance against other image extractors, or larger-image recovery.