PagePulse - Know what a book will do to you before you open it.
by•
PagePulse is a reading intelligence app for iOS and Android. Scan any bookshelf, see the emotional fingerprint of every book, understand what your library says about you.
Replies
Best
Maker
📌
Book discovery is broken, and the cost is invisible.
Every reader has bought a book they abandoned. Every avid reader has dozens. The economics of reading work against discovery, a book demands fifteen hours of your life, and you decide whether to commit it based on a cover, a blurb, and a friend's hot take.
PagePulse reads your bookshelf back to you.
Point your camera at any shelf. Gemini Vision reads every spine in a single multimodal call. Each detected book gets a five-emotion fingerprint, chapter by chapter.
Report
chapter-by-chapter emotional fingerprint is a cool idea but seems like it could work against itself - part of what makes a gut-punch chapter land is not seeing it coming. do you show the full chapter breakdown upfront, or is there a spoiler-safe mode that just gives an overall shape (e.g. "slow build, brutal back half") without mapping specific chapters to specific emotions?
Report
Maker
@galdayan Really good question, and honestly the tension you are naming is the one we argue about internally the most. Right now the chapter-by-chapter view is opt-in, not the default surface. The default read is the overall shape, a single fingerprint plus a one-line verdict like "slow build, brutal back half", exactly the framing you described. The per-chapter breakdown lives one tap deeper, behind a "show emotional arc" toggle, so a gut-punch chapter is not spoiled unless you go looking. Open question we are still working through: whether that toggle should warn you before revealing, or whether crossing the tap is enough consent. Leaning toward a soft warning on first use. Curious what you would want.
Report
Scanned a small stack on my nightstand and the emotional read on each book was weirdly accurate, like it picked up on the vibe of a slow-burn novel I couldn't put down. Actually made me want to dig through the rest of my shelves to see what it says about me.
Report
Maker
@hakanaknto2q This is the exact loop we designed for, scan a stack, feel seen, then want to know what the rest of the shelves are hiding. That "slow-burn novel I could not put down" pattern is one of the tags the model is best at picking up. If you do the full-shelf scan, tell us if any book got read completely wrong. That is the signal we are tuning against.
Replies
chapter-by-chapter emotional fingerprint is a cool idea but seems like it could work against itself - part of what makes a gut-punch chapter land is not seeing it coming. do you show the full chapter breakdown upfront, or is there a spoiler-safe mode that just gives an overall shape (e.g. "slow build, brutal back half") without mapping specific chapters to specific emotions?
@galdayan Really good question, and honestly the tension you are naming is the one we argue about internally the most. Right now the chapter-by-chapter view is opt-in, not the default surface. The default read is the overall shape, a single fingerprint plus a one-line verdict like "slow build, brutal back half", exactly the framing you described. The per-chapter breakdown lives one tap deeper, behind a "show emotional arc" toggle, so a gut-punch chapter is not spoiled unless you go looking. Open question we are still working through: whether that toggle should warn you before revealing, or whether crossing the tap is enough consent. Leaning toward a soft warning on first use. Curious what you would want.
Scanned a small stack on my nightstand and the emotional read on each book was weirdly accurate, like it picked up on the vibe of a slow-burn novel I couldn't put down. Actually made me want to dig through the rest of my shelves to see what it says about me.
@hakanaknto2q This is the exact loop we designed for, scan a stack, feel seen, then want to know what the rest of the shelves are hiding. That "slow-burn novel I could not put down" pattern is one of the tags the model is best at picking up. If you do the full-shelf scan, tell us if any book got read completely wrong. That is the signal we are tuning against.