paperless-classify

What problem does it solve?

It helps you quickly triage and organize Paperless-ngx documents by turning raw OCR text into consistent metadata, so you stop manually guessing correspondents, types, tags, and titles.

Core Features & Use Cases

• Inbox-first classification workflow: Review documents that still carry the inbox tag and classify them with human-visible context for accuracy.
• Metadata assignment with taxonomy alignment: Assign correspondent, document type (finite set), tags, and a normalized title using your existing taxonomy conventions.
• Bulk application + safety rails: Apply classifier decisions in batches from structured stdin, remove the inbox tag on update, and prompt for creation only when taxonomy gaps require it (with correspondent always created automatically).
• Compact-to-full content strategy: Use brief (compact content) for most docs, then selectively re-run with full content for ambiguous cases to improve classification quality.

Quick Start

Use the paperless-classify skill to review and classify your inbox documents by running the classify inbox flow.