Private by default
No account, database, analytics or shared notes.
Product evolution
Book discovery without a data account
Bookie began as a small search-and-save CRUD exercise. This version keeps that useful core while removing the shared anonymous database and clarifying every state.
What changed
A smaller architecture with a clearer promise
The original application combined a React client, Express server, MongoDB and a browser-delivered Google Books key. It had no account boundary, so a global “Recommended Books” list could be modified by any visitor.
Bookie Reading Lab instead treats the shelf as personal device data. A small Vercel Function protects the optional Google Books key, while curated pages remain fast and useful without live search.
Optional live search
Metadata search is a separate, bounded capability.
Purposeful shelf states
Want to read, Reading, Finished and Recommend.
Portable data
Export and import a readable JSON file.
Architecture
Static pages plus two read-only Functions
The build generates crawlable HTML and local assets. The only server operations are health reporting and Google Books metadata search.
Generated pagesBrowser-local shelfOptional search FunctionGoogle Books public metadata
Start reading
Choose the next useful step
Search widely, follow a curated path or continue a book already on your shelf.