Tanuki is a self-hosted media vault for mixed libraries that include videos, images, manga, comics and source downloads. It combines scanning, playback, reading, metadata management, downloads, auto-tagging, duplicate detection and collections in a Docker-based stack.
Tanuki is designed for self-hosted deployments with a shared library model, background workers, and persistent storage for media, database data and generated assets.
- Recursive scan of
/media - Automatic media type detection for common video, image and archive formats
- Thumbnail generation for local media
- Metadata editing in the browser
- Optional delete from database only or from disk as well
Supported library media types:
videoimagemangacomicdoujinshi
Supported archive reader formats:
.cbz.zip.cbr.rar
- Video playback with resume support
- Manga and comic reader with single-page, double-page, scroll and RTL modes
- Persistent reader and player preferences in the browser
- Queue-based download manager
- Scheduled downloads
- Batch URL submission
- Automatic organize and follow-up scan after completion
- Site-specific connectors plus
yt-dlp,gallery-dland direct HTTP fallback
Currently supported public connectors:
hentai0.comdoujins.comrule34.artdanbooru.donmai.ussafebooru.orggelbooru.com
- Tag search and tag-based filtering
- Reverse-image auto-tagging
- Duplicate detection via perceptual hash
- Manual collections
- Smart collections based on type, title, tag, favorite flag and minimum rating
- Multi-user authentication
- Admin-only user management
- Admin-only plugin management
- Docker
- Docker Compose
git clone https://github.com/nacl-dev/Tanuki.git
cd Tanuki
cp .env.example .env
docker compose up -d --buildOpen the application:
On first start, the default Compose setup creates media/ and inbox/ automatically. If SECRET_KEY is left empty, Tanuki generates one and stores it persistently. The first registered user becomes admin.
If you want to skip local image builds, you can use the published GitHub Container Registry images instead:
git clone https://github.com/nacl-dev/Tanuki.git
cd Tanuki
cp .env.example .env
docker compose -f docker-compose.ghcr.yml up -dThis uses the published tanuki-app, tanuki-worker and tanuki-downloader images. By default it pulls latest. To pin a specific release, set IMAGE_TAG in .env before starting the stack.
The stack works with sensible defaults. These are the settings you are most likely to change. See .env.example for the full list.
| Variable | Default | Purpose |
|---|---|---|
PORT |
8080 |
HTTP port exposed by the app container |
MEDIA_PATH |
/media |
library root |
INBOX_PATH |
/inbox |
import root |
DOWNLOADS_PATH |
/media |
allowed download root |
SECRET_KEY |
auto-generated if empty | application secret, minimum 32 characters |
REGISTRATION_ENABLED |
true |
allow self-registration |
SCAN_INTERVAL |
300 |
automatic scan interval in seconds |
DOWNLOADER_COOKIES_FILE |
empty | optional Netscape cookies.txt path |
SAUCENAO_API_KEY |
empty | SauceNAO API key for auto-tagging |
BASE_URL |
/ |
frontend base path when deploying behind a reverse proxy prefix |
If you access Tanuki behind a reverse proxy subpath, set BASE_URL to that prefix before building the frontend.
Tanuki uses a shared-library model with user-specific areas where it makes sense:
- media files and tags are shared across the instance
- collections are user-scoped
- download jobs are user-scoped
- download schedules are user-scoped
- runtime and path details are visible only to admins
- plugins are admin-only
Tanuki does not currently provide strict per-user library isolation.
- keep the
configvolume persistent if you use the auto-generated secret - put the stack behind a reverse proxy with TLS
- keep
.envout of version control - back up PostgreSQL and the
media/volume regularly - stricter source sites may still require browser-exported cookies
Tanuki is source-available for noncommercial use. Commercial use is not permitted under this license.