Docspell
A self-hosted document management system that uses OCR and machine learning to make scans and emails searchable, with Docker support and an AGPL-3.0 license.
Docspell is a self-hosted document organizer for households and small organizations that need to find files collected from scanners, email and other sources. It uses extracted text to help reduce manual labeling.
- OCR and search: Tesseract reads scans and images so you can search their text, alongside document names and notes, and filter results by tags or correspondent.
- Metadata suggestions: Machine learning and NLP help identify correspondents, tags and dates, while address-book matching can automatically attach matches to incoming documents.
- File handling: Convert imported files to PDF while retaining untouched originals, create searchable PDFs from scanned images, and import the contents of ZIP archives and EML email files.
- Shared collections: Multiple users can work with the same files while keeping individual settings, using a mobile-friendly browser interface or an Android app for uploads.
- Email and notifications: Import IMAP mailboxes on a schedule, send documents through SMTP, and receive event or recurring-search notifications through email, Matrix or Gotify.
Docspell runs on your own server and supports Docker deployment. It's open source under GNU Affero General Public License v3.0 (AGPL-3.0). Its text analysis uses Stanford CoreNLP, and a REST/HTTP API lets other software access documents and functions.
