About Shellf
I’m Jayson Robinson, and Shellf is the reading app I’m building. I’m the only person who works on it. This page is the short version of what it is, why I’m building it, and what it does that other reading apps don’t.
The product
Shellf is a book tracking app. It’s on Android today, with iOS arriving mid-2026, and shellf.app as the companion website.
It does what you’d expect a tracker to do: log books, organise shelves, rate what you read, see your reading stats. Most book apps stop there. Shellf does two more things.
The first is recommendations. Most apps push the bestseller list. Shellf builds a model of your specific taste from your library and ratings, recommends books that match it, and writes a reason for every pick. The second is reader insights: the threads that link your favourites, what your taste actually looks like, how it changes over time. Things you can’t get from a “books read this year” count.
That’s a slightly different take from most trackers, not an opposite one. Shellf still does the tracking. It just doesn’t stop there.
The maker
I’m Jayson Robinson. I work on Shellf alone. There’s no team. No investors. No roadmap I have to defend in a board meeting.
Before this I spent over a decade in product management and growth marketing across a mix of startups and scale-ups. Shellf is the first major product I’ve built myself.
I work from a desk in southeast London with my wife, daughter and two dogs. I read a mix of fiction (mostly classics and popular literary) and non-fiction (mostly history, philosophy and social science).
My favourite book of all time is The Count of Monte Cristo by Alexandre Dumas.
Why I’m building it
My wife made me 😅. She’d been badgering me for over a year to build her something that gave her better recommendations and better stats than the existing apps (mostly StoryGraph and Goodreads). She’s a voracious reader.
As well as working in tech, I’m also an entrepreneur. So I did some market research and thought it was a market worth investing time in — the existing apps weren’t solving the problem well enough for her, or for others I spoke to.
That gave me a problem worth working on. The dominant app is owned by a company that hasn’t materially improved it in years. The challengers tend to solve for one slice of reader behaviour and leave the rest. And it’s cheaper and easier to build software than it’s ever been, so I thought I could make a better product, more cheaply, than what’s already out there.
I’m building Shellf around three bets:
Taste is more specific than genre.
An app should be able to model what you actually read for, not which shelf the bookshop puts you on.
Reading data has more to say than “47 books this year.”
The patterns are in there. Most apps don’t bother surfacing them.
Staying small is a feature, not a limitation.
No social feed, no ads, no growth-at-all-costs. The constraints shape the product.
These are bets, not conclusions. Shellf is small — a few hundred users so far, not a few million. I haven’t earned the right to call any of them proven yet. I’d rather be open about that than pretend otherwise.
How it works
The product is built around three things. None of them are novel on their own. The bet is that doing each one well makes them reinforce each other.
01
Library
Every book you've read, are reading, or want to read. Per-trait rating tags let you mark plot, dialogue, characters, and pacing on a single book separately, instead of flattening the whole experience into four stars.
02
Recommendations
An LLM and a text-embedding model build a profile of your taste from your library and ratings. Multiple retrieval strategies (similar books, cross-cluster picks, wildcards) keep it from echo-chambering. Every pick comes with a written reason.
03
Reader insights
Clusters and patterns across your library, surfaced as a taste fingerprint. The threads that link your favourites; how your reading has changed over time.
Our principles
Shellf is built on five principles. Each came out of conversations with the readers I’m building the app for.
Built for serious readers
Shellf is for serious readers. Bibliophiles. People with hundreds of books in their library, who finish 30+ a year, and have strong opinions about what they read. My wife Helen is the prototype user; she reads more than I do and was the first person to use the app properly. A handful of friends in the same bracket are my initial user base and informal research panel. Most product decisions get checked against what they'd actually use.
Private by default
Shellf isn't a social app. There's no feed, no followers, no public profile. When you log a book, rate it, or write a note, no one else sees it. Public list-sharing is an opt-in feature for the people who want it; private is the default. Almost every early user I talked to said the same thing — they didn't want their reading life to be performative.
Library flexibility
Serious readers organise their books in idiosyncratic ways. By mood, by physical shelf, by historical era, by which ones they want to re-read, by which ones their mum recommended. Shellf is built around that. Custom shelves, flexible tags, multiple ways to slice your library. The early users I talk to are picky about how they organise, and the app reflects that rather than forcing a one-size-fits-all default.
Ratings with depth
Recommendations are only as good as the data behind them, and that data has two dimensions: breadth (how many books you've logged and rated) and depth (what you've said about each one — an overall rating, per-trait rating tags for plot, characters, dialogue and pacing, plus your notes). A four-star rating on its own tells the engine very little. The same rating with 'loved characters, hated pacing' and a 200-word note tells it a lot. The more you give the engine, the sharper the recommendations get.
Insight, not just stats
Helen's original ask was the most interesting one, and it's what pushed Shellf past being a tracker. She'd spent years half-finishing a particular kind of book before she realised what they had in common: none of them gave her a character to root for. Until then, she'd just thought she was bad at finishing those books — she didn't have a name for the pattern. Shellf's job is to make patterns like that easier to find. They're already in what you read; most apps just count the books.
Get in touch
X / Twitter
@jaysonrobinsonPersonal site
jaysonrobinson.comThe app
Shellf on Google PlayOr just install the thing: