Book Companions
Personalized book guides help readers revisit ideas that matter to them. AI-assisted additions and edits keep each companion useful.
Explore Book Companions ↗ (opens in a new tab)BookWard / Reading & discovery
Remember what you read. Discover what comes next.
What started as a folder of AI-written book summaries became a personal reading companion: a place to remember ideas, understand your taste, and discover beyond it.
BookWard.aiCreated & operated by PharettaLabs01 / Discover
It started with a forgotten book.
John likes bringing ideas from books into conversation. Remembering which book an idea came from, who wrote it, or the details that made it useful was another matter. Taking consistent reading notes had never become a habit.
His first solution was simple: create book summaries with AI and keep them in Google Drive for reference. As the folder grew, finding the right summary became its own problem. A small app seemed like a better home.
That personal workaround became the starting point for BookWard—and for one of its core features, Book Companions.
The question behind the product
What if your reading library helped you remember why a book mattered to you?
02 / Define
A companion shaped by the reader.
A Book Companion is an AI-written guide shaped by what BookWard knows about a reader: the books they like, the ones they do not, and the reasons behind those preferences. It focuses on ideas the reader may want to remember, and the reader can ask AI to add to or modify it.
That opened up the broader product. The same reading history and feedback could help organize a library, guide recommendations, and support reading goals.
The aim became a reading experience that grows with the person using it. Helping someone switch from an existing tracker also became part of the problem: their history should be a starting point, rather than something they have to rebuild by hand.
The product principle
Remember the reasons behind a reader’s taste—and leave room for that taste to change.
03 / Make
Connect memory, taste, and discovery.
BookIQ & mood-based discovery
Personal fit scores draw on reading preferences and feedback. Readers can also describe what they are in the mood for when looking for the next book.
See how recommendations work ↗ (opens in a new tab)“Prove me wrong”
A recommendation path deliberately challenges the fit score, suggesting books that may score lower but could still surprise the reader.
A library that comes with you
Import reading history from supported exports, spreadsheets, lists, or copied library pages from services such as Goodreads, Amazon, and Audible. AI helps extract book lists from messy text.
See how book-list imports work ↗ (opens in a new tab)Reading goals & personal shelves
Track reading, set goals, and create shelves organized around personal interests and preferences.
Explore the book tracker ↗ (opens in a new tab)From the bookshelf to the phone
Camera-based cover and barcode scanning make it easier to add books. Push notifications are supported across web and native mobile experiences.
Explore all BookWard features ↗ (opens in a new tab)
The hardest part / The book catalog
The AI needed a reliable world of books.
The biggest challenge was the underlying catalog. Books come from many countries and languages, and a single title can have multiple editions, publication dates, and inconsistent records.
John brought together multiple book-data repositories and worked through the practical problems: author accuracy, cover images, duplicate titles, edition matching, and original publication dates. Search also needed useful suggestions across a very large catalog, while book pages needed real previews where available.
Combining sources, cleaning data, and using AI assistance made the catalog workable. Keeping those choices cost-conscious was part of making the service accessible.
Product data / Public discovery
Put the catalog to work beyond the app.
BookWard also turns book data into public discovery pages, with monthly rankings, genre breakdowns, and chart history. Readers can explore before creating an account.
Top 100 Books
A monthly view of what readers are picking up now, spanning all-time favorites and current titles. BookWard Score rankings, chart history, and month-to-month movement make the trends visible.
Explore the Top 100 ↗ (opens in a new tab)Rising Books
A companion chart focused on newer releases, with genre breakdowns that help readers explore books gaining attention.
Explore Rising Books ↗ (opens in a new tab)Public popularity rankings use BookWard Score. Personalized recommendations use the reader’s BookIQ fit score.
Technology & implementation
The tools behind BookWard.
AI personalization sits alongside catalog data, hosted infrastructure, and capabilities available on a reader’s device.
- Supabase
- Backend platform
- Anthropic API
- AI features within the reading experience
- Resend
- Email delivery
- Netlify
- Hosting and deployment
- Porkbun
- Domain management
- Book-data sources
- Multiple repositories supplying catalog information, with data cleaning and record matching
- Device cameras
- Book-cover and barcode scanning on iPhone and Android, across web and native app experiences
- Push notifications
- Notifications across web and native mobile experiences
Architecture detail
Catalog sources, mobile capabilities, and AI services work together to support the reading experience.
04 / Launch
Make it easy to bring your books along.
BookWard grew from John’s personal reference tool into a product he builds and operates. The experience spans the web and native mobile apps, with camera access and notifications extending what readers can do away from a desktop.
Imports help reduce the work of starting over in a new tracker. Readers can bring an existing library, organize it, and begin giving the feedback that shapes their experience.
BookWard is free during beta. Keeping costs manageable has been part of the product’s development.
Explore BookWard ↗ (opens in a new tab)05 / Learn & improve
A good match can still be a surprise.
Personalization creates a useful tension: learning someone’s taste can help find better matches, but it can also keep recommending more of the same.
“Prove me wrong” is BookWard’s response. It gives lower-scoring books a chance to challenge the model’s assumptions and the reader’s expectations. The score is a guide to explore, rather than a boundary around what someone might enjoy.
The public rankings also connect product development with discovery: the same catalog supports useful pages that people can find and explore outside the app.
As readers rate books and explain what worked for them, the experience is designed to keep adapting. Improved recall and broader discovery are intentions behind the product; this story does not present measured reading outcomes.
The PharettaLabs connection
Start with a personal problem. Build on reliable data. Make room to discover something unexpected.
Have a problem worth solving?
Let’s open up
what’s possible.
Tell us what you’re imagining ↗Bring this kind of thinking to your business
Custom web apps ↗Spreadsheets to software ↗Organic growth ↗