Paperless-ngx: Office files, tags and email setup

This Paperless-ngx tutorial extends an existing Unraid setup with Office document conversion and automated email importing. The presenter installs Gotenberg and Apache Tika, then enables their endpoints through the Paperless container variables. When the Tika template fails to pull its image, he replaces its repository with the official Apache Tika image on Docker Hub.

A DOCX upload demonstrates the conversion workflow: Paperless keeps the original Word file, Gotenberg creates a PDF for preview, and Tika extracts searchable text. The presenter notes that enabling Tika does not automatically reprocess previously uploaded documents.

The tagging walkthrough covers colors, parent and child tags, and keyword matching. An "any" rule assigns a server tag when a document contains the chosen keyword; "all" requires every specified word. Existing documents need manual attention in the demonstrated workflow. The presenter suggests building up tagged examples before using automatic learning, with 20 or 30 documents as his rough guideline.

For email intake, he connects a test Gmail account through IMAP on port 993 with an app password. A sender filter imports attachments and adds an email tag, with a required maximum age set to 30 days. Manually triggering mail processing imports a Word attachment and applies both email and keyword tags. Local AI automation is reserved for a later installment.