Folders and tags: organising a scannable library
The Documents list in Paper Scan supports both nested folders and multi-tags. The two models are not redundant: folders answer 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), and 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), and 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".
Recommended workflow
- Scan directly into the appropriate folder. Use the existing hierarchy rather than creating new folders during capture.
- Apply tags that cross folder boundaries. A document in
2026 / Q2 / Acme / contractsmay also carryurgentandq2-renewalso it can be retrieved across all clients at renewal time. - Sweep the tag view periodically to file documents that were scanned in a hurry.
The combined search supports both folder and tag filters together, which is the main 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.
For libraries under 100 documents that fall into obvious categories, folders alone are sufficient until the library grows large enough that cross-cutting searches become common, typically around 200 documents.