Teedy

Last reviewed Sep 13, 2026 · 863 words

Pick Teedy when the question is "who is allowed to see this document", and pick Paperless-ngx when the question is "where did I put this document". That is the whole split. Teedy, formerly Sismics Docs, is a 2,563-star document management system built for groups: users, groups, per-document access lists, review workflows, and LDAP login. Paperless-ngx is ten times more popular, has a far better consumption pipeline, and is designed around one household's paper. Both do OCR and full-text search. Only one of them has real permissions.

The team features are the product

A Teedy document is a container with a title, tags, language, a set of attached files, and an access control list. Share it with a user, a group, or a public link; give read or write; put it into a workflow where named people must validate it in order. Groups nest, so "Finance" can inherit from "Staff". Tags are hierarchical and carry their own ACLs, meaning a tag can act like a folder that only some people see. Add two-factor login, an audit log per document, and LDAP so the accounts come from the directory you already run, and you have something a 10-person office or a small nonprofit can use to hold contracts, HR files and board minutes without a shared drive free-for-all.

For the single-user home archive none of this matters, and the document management category has better answers there. The Paperless-ngx guide is where most readers of this page should go next.

The setup that holds up

The official image is sismics/docs, and it bundles Tesseract, so OCR works on day one:

services:
  teedy:
    image: sismics/docs:latest
    ports:
      - "8080:8080"
    environment:
      - DOCS_BASE_URL=https://docs.example.com
      - [email protected]
      - DOCS_ADMIN_PASSWORD_INIT=change-me-now
    volumes:
      - ./teedy-data:/data
    restart: unless-stopped

Three things matter here. DOCS_BASE_URL must be the public URL because share links and email notifications are built from it. The _INIT variables set the admin account on first boot only; without them the default is admin/admin, which needs changing before the container sees a network. And the /data volume holds everything, including an embedded H2 database by default. H2 is fine for small teams, but the image also supports PostgreSQL through DATABASE_URL, DATABASE_USER and DATABASE_PASSWORD variables, and I would switch to Postgres before the archive passes a few thousand documents; backups and recovery are far more predictable.

Budget 1 GB of RAM and a slow first index

Teedy runs on the JVM, and the catalogue's 1,024 MB minimum is honest. Idle it sits around 500 to 700 MB; during OCR of a 40-page scan it will use the rest and a full core for a minute or two. Do not put it on the same 2 GB box as three other Java services. OCR runs on upload, per language you select on the document, and you can add Tesseract language packs to the image if your paperwork is not in the bundled languages.

Full-text search covers OCR text, titles, tags and metadata, and it is quick once indexed. What Teedy lacks is the inbox pattern: there is no watched consumption folder where a scanner drops PDFs and the system files them. Uploads come through the web UI, the REST API, the Android app, or an email inbox if you configure SMTP and the inbound address. For a scanner workflow, a 10-line script against the API is the usual bridge, and that is a step Paperless-ngx never needs.

The interface is dated and that is fine

The UI is AngularJS from an earlier era and it looks it. Nothing is confusing, it is simply not pretty, and mobile use is via the separate Android app rather than a responsive layout. Development is steady rather than fast; the codebase is mature and releases arrive a few times a year with fixes and dependency bumps. The upside of a decade-old project is that the data model has not changed under anyone, and a backup from 2 years ago restores.

If a modern look matters and the team features are still needed, Docspell is the closest alternative with a fresher interface and a proper consumption pipeline, though it lacks Teedy's workflows. For an Evernote replacement at the personal level, neither is the right tool.

What I'd do

For a family archive: Paperless-ngx, no contest. For a small organisation with more than 3 people who need different access to different files: Teedy on PostgreSQL, 2 GB of RAM reserved, LDAP or the built-in accounts, behind a reverse proxy with the base URL set correctly, and a nightly pg_dump plus a tar of /data. Write the little API script for the scanner, set up one review workflow for anything involving money, and leave the rest of the features until someone asks for them. Teedy will not delight anyone, but it will keep the right documents in front of the right people, which is the only thing a team DMS has to do.