Skip to content
Paper Scanner
en
Back to blog

Folders and tags: organising a scannable library

Paper Scanner supports nested folders and many-to-many tags in the same library. The two models answer different questions; this post explains when and how to combine them.

P Paper Scanner Team · · 2 min read

The Documents list in Paper Scanner supports both nested folders and multi-tags. The two models are not redundant: folders answer the question where does this document live, while tags answer what is this document. Used together, they make a library of several hundred documents findable without manual searching.

Folder model

Folders form a tree. Each document lives in exactly one folder. Folders work best for dimensions that are stable, hierarchical, and one-per-document:

  • Time-based archives (year / quarter / month)
  • Project structures (client / project / phase)
  • Physical locations (house / boiler / manuals)
  • Single-relationship accounts and vendors

Folder hierarchies are recommended to stay at four levels or fewer. Deeper nesting usually indicates that two dimensions are being encoded into a single tree, which makes retrieval brittle.

Tag model

Tags are flat and additive. A document may carry any number of tags. Tags answer cross-cutting questions that do not fit a single folder:

  • Document type (receipt, contract, manual, warranty)
  • Action status (urgent, to-review, to-file)
  • Time-bound grouping (tax-2024, tax-2025)
  • Source (scanned-here, imported)

The tag view groups all matching documents across folders. This is the primary mechanism for answering questions like “show me every receipt from last quarter” or “everything I scanned this week”.

  1. Scan directly into the appropriate folder. Use the existing hierarchy rather than creating new folders during capture.
  2. Apply tags that cross folder boundaries. A document in 2026 / Q2 / Acme / contracts may also carry urgent and q2-renewal so it can be retrieved across all clients at renewal time.
  3. Sweep tag views periodically to file documents that were scanned in a hurry.

The combined search supports both folder and tag filters together, which is the primary advantage of using both models.

Patterns to avoid

  • Excessive tagging. Two to three tags per document is the practical limit. More tags dilute their meaning and reduce search precision.
  • Years in tags. Years belong in folders because they form a natural tree. Tags should describe what a document is, not when it happened.
  • Deep folder nesting. Four levels is the practical maximum. Past that, split out a dimension as a tag.
  • Renaming folders. Folder paths are referenced by share links, exports, and sync. Rename only when necessary; prefer tagging for re-classification.

When folders alone are sufficient

For libraries under 100 documents that fall into obvious categories, tags add overhead without proportionate benefit. Folders alone are sufficient until the library grows large enough that cross-cutting searches become common — typically around 200 documents.

Related posts