Papermerge — Self-Hosted Document Management for Digital Archives

Introduction

Papermerge is a self-hosted document management system designed for scanned documents and PDFs. It applies OCR to turn image-based documents into searchable text and organizes them in a hierarchical folder structure with tags and metadata.

What Papermerge Does

• Applies OCR to scanned documents and image-based PDFs automatically
• Provides full-text search across all ingested documents
• Organizes files in a hierarchical folder tree with drag-and-drop
• Supports tagging, metadata fields, and document versioning
• Offers a REST API for integration with external workflows

Architecture Overview

Papermerge is a Python/Django application with a React-based frontend. OCR is handled by Tesseract, which runs as a background worker processing uploaded documents. The system stores metadata in PostgreSQL and files on the local filesystem or S3-compatible storage. A task queue (Redis + Celery) manages asynchronous OCR and indexing jobs.

Self-Hosting & Configuration

• Deploy via Docker Compose with the official images for the app and worker
• Configure OCR languages by setting the PAPERMERGE_OCR__DEFAULT_LANGUAGE variable
• Use PostgreSQL as the database backend for production deployments
• Mount a persistent volume for /media to preserve uploaded documents
• Optionally configure S3-compatible storage for document files

Key Features

• Automatic OCR on upload converts scanned pages into searchable text
• Full-text search powered by a search index across all document content
• Dual panel document viewer for side-by-side browsing and organization
• Role-based access control with user and group permissions
• REST API enables programmatic upload, search, and metadata management

Comparison with Similar Tools

• Paperless-ngx — focuses on consumption and auto-tagging; Papermerge offers a richer folder hierarchy and dual-panel UI
• Mayan EDMS — enterprise DMS with workflows; Papermerge is lighter and faster to deploy
• Docspell — Scala-based with NLP tagging; Papermerge uses standard OCR and is Python-native
• Teedy — Java-based lightweight DMS; Papermerge has stronger OCR integration
• Alfresco — enterprise-grade, heavy; Papermerge targets home and small-team use

FAQ

Q: Which OCR languages are supported? A: Any language supported by Tesseract. Set the default language via environment variables and install additional language packs as needed.

Q: Can I import existing PDF files in bulk? A: Yes. Place files in the import directory or use the REST API for batch uploads.

Q: Does Papermerge handle multi-page documents? A: Yes. Multi-page PDFs are fully supported with per-page OCR and thumbnail generation.

Q: What are the minimum system requirements? A: 2 GB RAM and 2 CPU cores are recommended for comfortable OCR processing speeds.

Sources

• https://github.com/ciur/papermerge
• https://docs.papermerge.io