monloader

monloader is the downloader that feeds monbooru. Paste the URL of a booru post, a pool, a tag search, an artist page, or a direct image link; monloader fetches the files and their per-post metadata, maps that metadata onto monbooru’s data model, and pushes each file into a monbooru gallery over its API.

The actual scraping is done by gallery-dl, so any site gallery-dl supports works. monloader adds the download queue, the metadata mapping, and the push into monbooru, with curated profiles that sharpen the result for 50-plus known sites.

It also runs in reverse: for a file monbooru already holds, monloader can find tags by file hash and by image similarity (IQDB, SauceNAO) across the boorus, and optionally from a local copy of the Hydrus Public Tag Repository, then fold them back into the image. See reverse lookup.

The two apps stay separate on purpose: monbooru holds your files and needs no internet access, while monloader has internet access but never touches the gallery files - everything flows through monbooru’s API. monloader is intended for your local network; its UI is not designed to be exposed to the public internet.