Discover · Save · Rank
Find a story that stays with you.
Search novels by title, author, tag, or premise. Keep your reading list and personal tier list in one place.
Personal catalogue
Library
Keep reading states current and review every tentative tier placement.
Taste model
Tier list
Drag titles between rows or reorder within a row. Every direct move becomes ranking feedback.
Inside NoveList
From a search to your own tier list.
Catalogue search finds stories. Your reading history shapes tentative placements for the books you save.
Explore
Find a book
Build the catalogue index
Titles, aliases, authors, descriptions, and tags become searchable records. The local sentence encoder turns each record into a 384-number vector.
Search by words or meaning
The same encoder processes your query. FAISS finds nearby vectors while exact titles and matching metadata contribute keyword signals.
Narrow the results
Selected tags filter the ranked results. Guests see the catalogue preview; signing in opens the full collection.
Search model all-MiniLM-L6-v2 runs in this service. The search index is rebuilt when catalogue metadata changes.
Organize
Make the list yours
Save what you read
Reading status, ratings, and tier choices are stored in your account's library.
Get a tentative tier
Your ratings and direct tier placements form a weighted baseline. Books sharing tags with those choices draw on those signals; reading status adjusts their weight.
Refine it by moving books
Moving, rating, or confirming a book adds feedback. Your chosen tiers stay in place while tentative placements for saved books update.
Tier logic A deterministic calculation in your library, specific to your own reading history.
Reading patterns
Analytics
A visual summary of your library, progress, taste signals, and local footprint.
Shared taste
Community
Build themed collections, follow tier lists, and talk about the novels in them.
Identity and security
Profile
Your public reading summary and private account controls are kept separate.
Import reading list
Upload a CSV with a title column and optional status column. Titles and known aliases are matched.
Account recovery
Replace your one-time recovery code. The previous code stops working immediately.
Metadata operations
Catalogue ingestion
Collection jobs rotate through later pages, skip complete records, and revisit incomplete metadata.
Questions
How NoveList works
How are tentative tiers chosen?
Direct placements and ratings establish your baseline. Shared tags and the title’s reading state adjust that estimate. Dropped titles start conservatively.
What teaches the ranking model?
Rating, dragging, assigning a tier, or confirming a suggestion records feedback. Later predictions use those choices, with more weight on completed and dropped titles.
Why might a cover be missing?
A source may not publish one, may block hotlinking, or may have moved the image. NoveList tries its local cache, a safe proxy, and the source URL before showing a placeholder.
Can readers add novels?
Readers can request a title. Administrators can run exact-title discovery or broader collection jobs so the shared catalogue stays controlled and deduplicated.