Settings

Configure Paperless-NGX connection and deduplication parameters.

Paperless-NGX Connection

Deduplication Parameters

Documents scoring above this threshold are grouped as duplicates.

Confidence Weights

Sum: 100/100
Jaccard Fuzzy

Discriminative Penalty

70%

Low (0-30%): Minimal impact. Template documents (e.g. monthly invoices with different dates) may still appear as duplicates.

Medium (40-70%): Recommended. Catches most template-based false positives while keeping true duplicates intact.

High (80-100%): Aggressive. Documents with any differing structured data are heavily penalized. May over-penalize minor OCR differences in dates or amounts.

System Information

Database Path
./data/paperless-ngx-dedupe.db
Total Documents
3,873
Pending Groups
384

Backup & Restore

Export Configuration

Download a backup of all settings and dedup parameters.

Download Backup

Import Configuration

Restore settings from a previously exported backup file.