Papermerge DMS (papermerge-core) is the open-source core of Papermerge, a document management system in Python that OCRs scanned documents, indexes them, and makes them searchable for anyone who needs a long-term digital archive rather than a folder full of loose files.

What it is

Papermerge is a document management system built for scanned documents, the kind of material often called a digital archive. It extracts text from scans with OCR, indexes that text, and prepares it for full text search. The interface borrows the look and feel of a modern desktop file browser, with a dual panel document browser, drag and drop, hierarchical folders, colored tags, and page-level management. It has been in development for nine years and lives in the Python ecosystem, distributed as an Apache-2.0 licensed project.

The concrete problem it replaces is the combination of a scanner output folder and a separate desktop search tool. Instead of dropping PDF, TIFF, JPEG, or PNG files into a shared drive and hoping filenames stay meaningful, Papermerge stores the documents themselves, keeps version history, attaches tags and typed metadata, and exposes everything through full text search over the OCRed content. It also replaces ad hoc sharing by email attachment with multi-user accounts, group ownership, and sharing of individual documents and folders between users or groups.

Key capabilities

• Web UI with a desktop-like dual panel document browser, drag and drop, tags, and hierarchical folders.
• OpenAPI compliant REST API for programmatic access, with Swagger docs served at /docs in a development server.
• OCR of documents plus full text search over the extracted text, and download of documents with the OCRed text overlay applied.
• Document versioning, so successive scans or revisions of the same record are preserved.
• Document Types (that is, categories) with Custom Fields for per-type metadata.
• Page management operations: delete, reorder, cut, move, and extract pages.
• Multi-user support with group ownership and sharing of documents and folders between users and groups, plus a UI available in multiple languages.
• pm command line tool for management commands, including pm search build to rebuild the search index from scratch and pm search stats for index statistics.

Who uses it and how

• Individuals and small offices that scan paper records and need long-term storage with text search rather than a folder tree on a NAS.
• Teams that need to distribute access to a shared archive, using group ownership and per-document or per-folder sharing instead of sending attachments around.
• Organizations with heterogeneous scan output, since the same system ingests PDF, JPEG, PNG, and TIFF files rather than requiring conversion first.
• Administrators and integrators who automate ingestion or reporting through the REST API, or who schedule index maintenance with pm search build and check it with pm search stats.
• Deployments that need localization, because the UI ships in multiple languages.

Getting started

The README starts with Docker: run docker run -p 8000:80 -e PAPERMERGE__SECURITY__SECRET_KEY=abc -e PAPERMERGE__AUTH__PASSWORD=123 papermerge/papermerge:3.5.3 for the most basic setup. For other deployment shapes, the documentation covers additional Docker Compose scenarios.

How it compares

This registry entry lists no paid products that Papermerge replaces, and the facts name no comparable open-source tools, so it stands alone in this registry. The only adjacent product mentioned in the README is Papermerge Cloud, the managed SaaS version by the same author, which is a commercial sibling rather than a named alternative.

When to use it — and when not to

A self-hoster must operate a PostgreSQL database (configured through PM_DB_URL) and a media root for document storage (PM_MEDIA_ROOT), so this is not a drop-in file share. The project is explicitly seeking maintainers: the author has shifted focus to Papermerge Cloud and opened issue #1318 for maintainer recruitment, with 36 open issues against the repository, which means prospective adopters should weigh long-term maintenance capacity before committing. Pick it if you want a self-hosted, Apache-2.0 OCR archive with a real API; avoid it if you need a vendor-backed SLA, or if nobody on the team can run PostgreSQL and keep an index rebuilt.

docker run -p 8000:80 \
    -e PAPERMERGE__SECURITY__SECRET_KEY=abc \
    -e PAPERMERGE__AUTH__PASSWORD=123 \
    papermerge/papermerge:3.5.3

$ uv sync

export PM_DB_URL=postgresql://coco:***@127.0.0.1:5432/pmgdb
export PM_MEDIA_ROOT=/home/eugen/var/pmgdata
export PM_API_PREFIX='/api'

$ uv run task server

cd frontend/

VITE_REMOTE_USER=admin
VITE_REMOTE_USER_ID=49e78737-7c6e-410f-ae27-315b04bdec69
VITE_REMOTE_GROUPS=admin
VITE_BASE_URL=http://localhost:8000
VITE_KEEP_UNUSED_DATA_FOR=1

cd frontend
yarn workspace ui dev

yarn workspaces list

    uv sync --group dev
    uv pytest

 uv sync --python 3.14

pm --help

pm search build

pm search stats