Downloading

last updated

Paste a URL into the add bar and press Enter: a single post, a pool, a tag search, an artist page, a manga, or a direct link to an image. Anything gallery-dl understands works. You can also send pages straight from your browser with the monsender extension.

Each URL becomes a job on the queue. monloader works out which posts it covers, downloads each one, translates its tags for monbooru and sends it to your gallery. Once monbooru has the file, monloader deletes its own copy; it keeps nothing.

Per-item outcomes#

Every post in a job ends with one of these:

Outcome Means
created a new image in monbooru
duplicate monbooru already had this file; any new tags were added to it
enriched a lookup or refresh added tags to an image monbooru already had
matched a PTR lookup found the image’s tags, and monbooru applied them
replaced an upgrade swapped in a better file
skipped_archive already downloaded before, so not again
skipped_unsupported a file type monbooru doesn’t take (audio, a document…)
failed something went wrong; the row says what

A failed item also carries an error code; the list is in API and development.

Big searches#

A tag search or an artist page can run to thousands of posts, so a job takes by default at most 200 of them (change that in Settings -> downloads). You can change this temporarily in the menu beside the add bar. It remembers your last pick. A job that hit its limit offers get next N for the next batch and get all for the rest. Pools and manga are exempt: you asked for the whole thing, so you get all of it.

Watch a tag search or an artist page and monloader checks it again on a schedule, so new posts keep landing in monbooru without you pasting the URL again. Click watch on a finished search in the queue, or paste the URL on the Watched page and pick how often to check it.

Retrying#

  • retry runs a job again. Posts it already downloaded are skipped (skipped_archive), unless the job had failed.
  • force download fetches everything again, skipped posts included: the way to get back a post whose image you deleted in monbooru.
  • Sending a single post again always fetches it again, and monbooru adds any new tags to the image it has. That’s how to refresh one post.
  • cancel stops a job; posts already done keep their outcome.

Stopping or restarting monloader cancels the jobs that hadn’t finished. They stay in the list so you can retry them, and one that was mid-download comes back marked interrupted, one click away from being queued again. Finished jobs clear themselves after a few days; the delays are in Settings -> downloads.

When a download fails and you want to know why, the [?] next to a finished item fetches that post again with gallery-dl’s debug log on, and shows you the log. Nothing is sent to monbooru.

Pause and resume#

The pause button on the queue (also in the monsender popup) holds all downloads: the one in progress finishes, then nothing new starts until you resume. Handy to line up a batch before letting it run. A restart un-pauses.

Pools and manga#

  • A booru pool comes in as a collection: each page is its own image with its own tags, in order.
  • A manga or comic is bundled into one .cbz archive, which monbooru opens in its reader. If any page fails to download, the job fails rather than sending half a book.
  • A manga title with chapters becomes one archive per chapter.

Some pages only link to files hosted elsewhere, like a forum thread. monloader follows those links and downloads what they point to.

Tagging by hash#

Paste a file’s md5 (on its own or as md5:<hash>) into the add bar and monloader imports the booru post that has that file, like any single post. It uses the same lookup chain that monbooru’s lookups use.