Upgrade StorageFeature request: automatically assign user-supplied tags (e.g. using AI)
I find tags useful for categorizing a library because they don't depend on mutually exclusive categorization by folder structure.
However, I personally struggle to apply tags consistently.
I would bet that an LLM-based model could consume a set of user-supplied tags and text data, and yield as output a set of relevant tags.
However, I personally struggle to apply tags consistently.
I would bet that an LLM-based model could consume a set of user-supplied tags and text data, and yield as output a set of relevant tags.
https://github.com/kinranlau/zotero_tag_recommender
While it doesn’t automatically add tags per se, it uses AI to suggest tags based on a paper’s title and abstract. It also takes your most frequent library tags into account so the suggestions better match your existing tagging style. You can then simply click to apply the ones you want.
It also supports multi-word tag autocomplete for manually adding tags. Unlike Zotero’s default first-word matching, you can type the beginning of any word in a tag (e.g. “gas”, “diff”, or “GDE”) and still match “gas diffusion electrode (GDE)”.
Hope it makes staying organized a bit easier!