Download images from the web into monbooru and backfill tags for the files already in it.
Paste a direct URL, an image or search from an online booru/gallery or any site supported by gallery-dl and monloader fetches the files and per-post metadata, maps it onto monbooru's data model, and pushes each file into a monbooru gallery over the REST API.
It also runs in reverse: for a file monbooru already holds, monloader looks up its tags by md5 and by image similarity (iqdb, SauceNAO) across the boorus and/or from an optional local copy of the Hydrus Public Tag Repository, and applies them to the image.
 |
 |
- Download images A direct link to an image, or anything else a gallery-dl extractor matches : booru post, pool, tag search, artist page.
- Pools and manga. A booru pool's pages import as an ordered collection; a manga or comic gallery bundles into a single
.cbzfor monbooru's reader. - Metadata mapped to monbooru. Tags by category (artist / character / copyright / meta / ...), rating, and source, normalized across booru families so tags land the way monbooru expects them.
- 50+ curated sites, plus a fallback. Profiles for the danbooru, e621, moebooru, gelbooru, and philomena families and a set of manga/comic sites; anything else gallery-dl supports still works through a generic fallback.
- Queue management. Each download shows whether it was added, was already in your library, got skipped, or failed so you can see at a glance what happened to every item. Re-submitting a search skips posts already fetched, while re-submitting a single post refreshes it, updating the image monbooru already holds to the source's current tags. Pause and resume downloads globally from the topbar (or the monsender popup) to queue up a batch before it starts.
- Find tags by file hash or image similarity. Paste an image's md5 to import the matching booru post, or have monbooru fetch tags for a file it already holds; monloader walks the sources you pick in order: exact md5 searches, plus iqdb and SauceNAO similarity lookups that find re-encoded copies from the image's thumbnail. An optional local copy of the Hydrus Public Tag Repository adds offline sha256 lookups.
monloader is a companion downloader for monbooru. It fetches images from the web and pushes them into your library :
flowchart LR
web["- Any booru or gallery supported by gallery-dl<br/>- Direct image URL"]
ptr["Hydrus<br/>Public Tag Repository"]
sender["<b>monsender</b><br/>browser extension"]
loader(["<b>monloader</b><br/>downloader"])
booru["<b>monbooru</b><br/>Your self-hosted booru"]
web -->|browse| sender
sender -->|send URL| loader
web -->|paste URL| loader
loader -->|push images, tags, source| booru
booru -.->|lookup by hash| loader
loader -.->|reverse lookup: md5 + similarity| web
ptr <-.->|"sync: sha256 tags + aliases + implications<br/>contribute back"| loader
classDef hub fill:#5c6bc0,stroke:#9fa8da,stroke-width:3px,color:#ffffff;
classDef tool fill:#16161c,stroke:#5c6bc0,stroke-width:1.5px,color:#e2e2e8;
classDef src fill:#16161c,stroke:#8888a0,stroke-width:1px,color:#8888a0;
class loader hub;
class sender,booru tool;
class web,ptr src;
- monbooru : self-hosted booru; organizes, tags, and serves your collection.
- monloader : this application; fetches files and per-post metadata (via gallery-dl) and pushes them into a monbooru gallery over the REST API; also answers monbooru's reverse lookups (boorus, IQDB/SauceNAO, Hydrus PTR) and source refetches.
- monsender : browser extension; sends the URL of the page you're currently browsing to monloader.
monloader ships in monbooru's docker/docker-compose.yml :
- Uncomment the
monloaderservice in monbooru's compose and start it:docker compose up -d monloader
- Open monloader at
http://localhost:8081, go to Settings -> pairing, and click connect to monbooru (the default api url in the monbooru section,http://monbooru:8080, works on the shared compose network). - In monbooru, approve the pairing request in its monloader settings. monloader stores the token monbooru issues; there are no keys to copy by hand.
- Back in monloader, pick a default gallery and save.
- Paste a URL into the command bar on the home screen and press Enter.
See the monloader documentation for help.
Intended for local network use. monloader's UI is not designed to be exposed to the public internet.
monloader is mostly a wrapper for gallery-dl, which does the actual scraping. monloader adds queue management, maps that output onto monbooru's data model, and pushes it over the API.
monloader's optional PTR tag lookup syncs against the Hydrus Public Tag Repository(PTR) via Hydrus Network's repository protocol. The tags, aliases, and implications it serves are the work of the hydrus community.