Reading Recommendations

Reading Recommendations

Description: Get personalized reading recommendations from your Calibre library using series progress, ratings, tags, reading history, and collaborative filtering across similar books.

Trigger Phrases

• "What should I read next?"
• "Recommend a book like [title]"
• "What's good in my unread [genre]?"
• "Suggest something from my TBR pile"
• "Find my next series to start"
• "What have I been neglecting?"

Tools

• query_books(sort="rating", unread=True, tags=[...], limit=20) — Find top-rated unread books in a genre.
• manage_analysis(operation="reading_stats") — Reading statistics: completion rate, genre distribution, pages read, author diversity.
• manage_analysis(operation="series_progress") — Series tracking: which series are started but unfinished, next-in-series ordering.
• manage_metadata(operation="show", book_id=...) — Get full metadata for a candidate book: description, rating, tags, series position.
• calibre_metadata_search(query="similar to [title]") — Semantic similarity search using LanceDB embeddings.

Workflow

1. Profile the reader: Call manage_analysis(operation="reading_stats") to understand reading patterns, preferred genres, and completion behavior.
2. Surface candidates: Use query_books(unread=True, sort="rating", tags=[preferred_genre]) to get top-rated unread books. Combine multiple tag filters for precision.
3. Series catch-up: Call manage_analysis(operation="series_progress") to find series with book 1 read but book 2+ unread — these are high-confidence recommendations.
4. Similarity match: For "like this book" queries, use calibre_metadata_search() with a descriptive query of the source book's themes.
5. Rank and present: Score candidates by (rating + recency + series_position). Present top 3-5 with reasoning: why this matches the reader's taste, what tags overlap.

Example

"Recommend my next read from unread fantasy with high ratings." → query_books(tags=["fantasy"], unread=True, sort="rating", limit=30) → filter top 5 → manage_analysis(operation="series_progress") to prioritize continuations → present with per-book reasoning.