Formerly Calibre Web Automated Book Downloader (CWABD)
A unified web interface for searching and downloading books from multiple sources - all in one place. Works out of the box with popular web sources, no configuration required. Add metadata providers, additional release sources, and download clients to create a single hub for building your digital library.
Fully standalone - no external dependencies required. Works great alongside library tools like Calibre-Web-Automated, Booklore or Audiobookshelf for automatic import.
- One-Stop Interface - A clean, modern UI to search, browse, and download from multiple sources in one place
- Real-Time Progress - Unified download queue with live status updates across all sources
- Two Search Modes:
- Direct Download - Search and download from popular web sources
- Universal Mode - Search metadata providers (Hardcover, Open Library) for richer book discovery and multi-source downloads (additional sources in development - coming soon!)
- Format Support - EPUB, MOBI, AZW3, FB2, DJVU, CBZ, CBR and more
- Cloudflare Bypass - Built-in bypasser for reliable access to protected sources
- PWA Support - Install as a mobile app for quick access
- Docker Deployment - Up and running in minutes
- Docker & Docker Compose
-
Download the docker-compose file:
curl -O https://raw.githubusercontent.com/calibrain/shelfmark/main/compose/stable/docker-compose.yml
-
Start the service:
docker compose up -d
Edge users: If you're tracking the main branch (
:devtag), use compose files fromcompose/edge/instead.
- Open
http://localhost:8084
That's it! Configure settings through the web interface as needed.
volumes:
- /your/config/path:/config # Config, database, and artwork cache directory
- /your/download/path:/books # Downloaded books
- /client/path:/client/path # Optional: For Torrent/Usenet downloads, match your client directory exactly. Tip: Point the download volume to your CWA or Booklore ingest folder for automatic import.
Note: CIFS shares require
nobrlmount option to avoid database lock errors.
Direct Download Mode (default)
- Works out of the box, no setup required
- Searches a huge library of books directly
- Returns downloadable releases immediately
Universal Mode
- Cleaner search results via metadata providers (Hardcover, Open Library)
- Aggregates releases from multiple configured sources
- Requires manual setup (API keys, additional sources)
Set the mode via Settings or SEARCH_MODE environment variable.
Environment variables work for initial setup and Docker deployments. They serve as defaults that can be overridden in the web interface.
| Variable | Description | Default |
|---|---|---|
FLASK_PORT |
Web interface port | 8084 |
INGEST_DIR |
Book download directory | /books |
TZ |
Container timezone | UTC |
PUID / PGID |
Runtime user/group ID (also supports legacy UID/GID) |
1000 / 1000 |
SEARCH_MODE |
direct or universal |
direct |
USING_TOR |
Enable Tor routing (requires NET_ADMIN capability) |
false |
Some of the additional options available in Settings:
- AA Donator Key - Use your paid account to skip Cloudflare challenges entirely and use faster, direct downloads
- Preferred source - Enable/disable and reorder available web sources for downloads
- Library Link - Add a link to your Calibre-Web or Booklore instance in the UI header
- Content Folders - Route fiction, non-fiction, comics, etc. to separate directories
- Network Resilience - Auto DNS rotation and mirror fallback when sources are unreachable. Custom proxy support (SOCK5 + HTTP/S), Tor routing.
- Format & Language - Filter downloads by preferred formats, languages and sorting order
- Metadata Providers - Configure API keys for Hardcover, Open Library, etc.
docker compose up -dThe full-featured image with built-in Cloudflare bypass.
Routes all traffic through Tor for enhanced privacy:
curl -O https://raw.githubusercontent.com/calibrain/shelfmark/main/compose/stable/docker-compose.tor.yml
docker compose -f docker-compose.tor.yml up -dNotes:
- Requires
NET_ADMINandNET_RAWcapabilities - Timezone is auto-detected from Tor exit node
- Custom DNS/proxy settings are ignored when Tor is active
A smaller image without the built-in Cloudflare bypasser. Ideal for:
- External bypassers - Already running FlareSolverr or ByParr for other services
- Fast downloads - Using fast download sources
- Alternative sources only - Exclusively using Prowlarr, IRC, or other sources
- Audiobooks - Downloading from Audiobookshelf-compatible sources
curl -O https://raw.githubusercontent.com/calibrain/shelfmark/main/compose/stable/docker-compose.lite.yml
docker compose -f docker-compose.lite.yml up -dIf you need Cloudflare bypass with the Lite image, configure an external resolver (FlareSolverr/ByParr) in Settings under the Cloudflare tab.
Authentication is optional but recommended for shared or exposed instances. Enable in Settings.
Alternative: If you're running Calibre-Web, you can reuse its user database by mounting it:
volumes:
- /path/to/calibre-web/app.db:/auth/app.db:roThe application exposes a health endpoint at /api/health (no authentication required). Add a health check to your compose:
healthcheck:
test: ["CMD", "curl", "-sf", "http://localhost:8084/api/health"]
interval: 30s
timeout: 30s
retries: 3Logs are available via:
docker logs <container-name>/var/log/shelfmark/inside the container (whenENABLE_LOGGING=true)
Log level is configurable via Settings or LOG_LEVEL environment variable.
# Frontend development
make install # Install dependencies
make dev # Start Vite dev server (localhost:5173)
make build # Production build
make typecheck # TypeScript checks
# Backend (Docker)
make up # Start backend via docker-compose.dev.yml
make down # Stop services
make refresh # Rebuild and restart
make restart # Restart containerThe frontend dev server proxies to the backend on port 8084.
┌─────────────────────────────────────────────────────────────┐
│ Web Interface │
│ (React + TypeScript + Vite) │
├─────────────────────────────────────────────────────────────┤
│ Flask Backend │
│ (REST API + WebSocket) │
├───────────────────┬─────────────────────┬───────────────────┤
│ Metadata Providers│ Download Queue │ Cloudflare │
│ │ & Orchestrator │ Bypass │
├───────────────────┼─────────────────────┼───────────────────┤
│ • Hardcover │ • Task scheduling │ • Internal │
│ • Open Library │ • Progress tracking │ • External │
│ │ • Retry logic │ (FlareSolverr) │
├───────────────────┴─────────────────────┴───────────────────┤
│ Release Sources │
├─────────────────────────────────────────────────────────────┤
│ • Direct Download (Anna's Archive → Libgen → Welib) │
├─────────────────────────────────────────────────────────────┤
│ Network Layer │
├─────────────────────────────────────────────────────────────┤
│ • Auto DNS rotation • Mirror failover • Resume support │
└─────────────────────────────────────────────────────────────┘
The backend uses a plugin architecture. Metadata providers and release sources register via decorators and are automatically discovered.
Contributions are welcome! Please file issues or submit pull requests on GitHub.
Note: Additional release sources and download clients are under active development. Want to add support for your favorite source? Check out the plugin architecture above and submit a PR!
MIT License - see LICENSE for details.
This tool can access various sources including those that might contain copyrighted material. Users are responsible for:
- Ensuring they have the right to download requested materials
- Respecting copyright laws and intellectual property rights
- Using the tool in compliance with their local regulations
Downloads are written atomically (via intermediate .crdownload files) to prevent partial files from being ingested. However, if your library tool (CWA, Booklore, Calibre) is actively scanning or importing, there's a small chance of race conditions. If you experience database errors or import failures, try pausing your library's auto-import during bulk downloads.
For issues or questions, please file an issue on GitHub.




