Search, download, organize, and read your periodicals
View automated tasks and their last execution times. You can manually trigger tasks as needed.
Monitor and manage your periodical downloads. Auto-refreshes every 5 seconds.
No downloads in queue
Tip: Go to Track Periodicals → Find Issues → Click Download to submit downloads
| Periodical | Status | Summary |
|---|
For finding and downloading periodical issues (Newsnab, RSS, etc.)
Downloads are automatically routed by provider type: Newsnab/RSS → SABnzbd/NZBGet, Torznab → qBittorrent, Internet Archive → built-in HTTP client.
Available tags: {category}, {title}, {year}, {month}, {day}, {volume}, {issue}, {language}
Extract text directly from PDF/EPUB files during import (fast, synchronous).
Use OCR on cover images for text extraction (slower, runs in background queue).
Automatically remove folders without importable files from downloads folder. Disable if downloads are being removed before processing.
Automatically remove empty folders from library after file reorganization. Generally safe to keep enabled.
Nested folder scanning in Downloads is always enabled.
Tools for reorganizing and maintaining your library files.
Clean up messy folder structures and reorganize files to match your current organization pattern. This will move files to proper locations and remove old directories.
Available tags: {category}, {title}, {year}, {month}, {day}, {volume}, {issue}, {language}
Higher values = stricter matching. Used for duplicate detection and title matching.
Publications with same title within this many days are considered duplicates.
Control how metadata from different sources (OCR, text scan, filename) is combined.
Drag to reorder. First = highest priority.
When multiple sources provide metadata, the system tries them in this order until one meets the confidence threshold.
For each metadata field (year, month, issue number, etc.), the system tries sources in priority order. A source is only used if its confidence score meets or exceeds its threshold.
Example: For a file "Magazine - Jan 2024.pdf":
Result: Year = 2024 (from filename, since OCR wasn't confident enough)
How confident OCR must be before using its values (default: 70%). OCR has per-field confidence scores based on text recognition quality.
Threshold for direct text extraction from native PDFs/EPUBs (default: 50%). Text scan is treated as 100% confident when available.
Threshold for filename parsing (default: 0%, always accept as fallback). Filename parsing is treated as 100% confident.
Fine-tune OCR confidence for specific metadata fields.
Years are critical - higher confidence recommended (default: 80%).
Month names are easier to OCR (default: 60%).
Issue numbers are moderately important (default: 75%).
Volume numbers are moderately important (default: 75%).
Configure download limits and concurrent downloads.
Number of times to retry a failed download before marking it as failed (default: 1).
Maximum number of downloads allowed system-wide across all tracked periodicals (default: 10).
Control how often scheduled tasks run (in seconds). These intervals match the tasks shown in the Tasks tab.
How often to search for and download new tracked issues (default: 1800 = 30 minutes).
How often to check download status and import completed files (default: 30 seconds).
How often to remove orphaned covers and generate missing ones (default: 86400 = 24 hours).
Configure adaptive search scheduling for tracked periodicals. Instead of searching ALL tracked periodicals every run (which would overwhelm providers), the system searches a small batch per run and adjusts intervals based on discovery success.
How many periodicals to search per auto-download run (default: 10). This does NOT search all periodicals—it searches a small batch and adapts intervals based on success. Higher values = more provider load but faster discovery. Lower values = gentler on providers but slower discovery.
How often to search periodicals that are actively finding new issues (default: 0.5 hours = 30 minutes).
Default search frequency for periodicals with 1-2 consecutive empty searches (default: 2 hours).
How often to search periodicals with 3-5 consecutive empty searches (default: 12 hours = twice daily).
How often to search very quiet periodicals with 6+ consecutive empty searches (default: 48 hours = every 2 days).
Configure DPI and quality for PDF cover extraction.
DPI for thumbnail/preview covers (lower = faster, smaller files).
DPI for high-quality covers (higher = better quality, larger files).
JPEG quality for thumbnail covers (1-100).
JPEG quality for high-quality covers (1-100).
Configure image preprocessing for better OCR accuracy.
Image width for OCR processing (smaller = faster, larger = more accurate).
Contrast boost for better text recognition (1.0-2.0, default: 1.5).
Noise reduction strength (higher = more smoothing, may blur text).
Sharpening kernel size (must be odd: 3, 5, 7, or 9).
Configure OCR processing performance and resource usage.
Number of parallel OCR workers (1-8). More workers = faster processing but higher CPU/memory usage.
Number of jobs to process per batch (1-20). Larger batches process more files at once.
Update your login credentials
Use this token to authenticate API requests. Treat it like a password and keep it secret.
Authorization: Bearer <token>
Restart the application to apply configuration changes
Restart the Curator application. This will briefly interrupt all operations.
Manage your library database
These actions cannot be undone. Use with caution.
Remove all library entries and tracking records from the database. Files on disk will remain untouched.
Caches NZB file content to avoid repeated provider hits on retries
When enabled, NZB files are fetched once and cached locally. Download retries use cached content without hitting the provider again.
This action cannot be undone. Use with caution.
Clear all cached NZB content. Future downloads will re-fetch from providers.