The genres, moods, and tags are getting a little chaotic. There are quite a few duplicates floating around (nonfiction, non-fiction etc), sometimes due to typos, and some overlap between categories too (genre: adventure, mood: adventurous, tag: adventure). This makes discovery harder for readers and also muddies up statistics.

A couple of ideas that might help:

1. LLM assistance. Basically an AI pass to merge duplicates and synonyms into canonical terms or summaries, without needing to share any user data with third-party services.

2. Curated controlled vocabulary. Swap free-text input for a well-thought-out predefined list, keeping things consistent from the start.

Either way, a bit more structure here would go a long way.