Building a Bookstore, Not a Dashboard
I built InkTree in ten hours. Then I spent a hundred making it worthy of the books inside it. Before and after a design overhaul driven by craft, not speed.
On reading, taste, and building tools for readers who care.
I built InkTree in ten hours. Then I spent a hundred making it worthy of the books inside it. Before and after a design overhaul driven by craft, not speed.
408 books on Goodreads. 300 rated 4 stars. Dostoevsky and Dungeon Crawler Carl, same number. The system is broken — and nobody's building a better one.
Building a cover art pipeline for 5,500 books using free APIs — placeholder detection, rate limits, and why Google Books sometimes returns a table of contents.
I thought building a Goodreads import would be simple. Then I discovered editions vs works, fragmented reviews, and why book cataloguing is a rabbit hole.
Star ratings compress everything a book does to you into one number. 13,000 dimensional reviews later, the data confirms what every reader already knows.
Three books, three completely different experiences, one number for all of them. I built InkTree because star ratings don't capture why a book matters.