The starting point
There is a particular kind of optimism in saving an article. You see a thoughtful headline, open a new tab, and tell yourself you’ll come back when there is more time. Then the tabs multiply. The bookmarks become a cupboard with no labels. The good ideas are technically safe, but practically lost.
I wanted a reading list that felt less like a queue and more like a quiet shelf: available when I need it, invisible when I don’t. So I started building one. The first decision was less about the interface and more about where the data should live.
Good tools should make it easier to return to your own attention.
Why local-first?
“Local-first” is a useful design constraint. It means the primary copy of your work lives on your device, so the tool remains useful without a connection and responsive without asking a distant server for permission.
For a reading list, that means links can be saved on a train, notes can be edited in a cabin, and the collection remains yours even if a service changes its pricing or disappears. Sync is still valuable — but it becomes a convenience rather than a dependency.
Your saved links and notes are portable, inspectable, and easy to export.
The essential actions work first, even when the network does not.
No feed, streaks, or clever pressure to turn curiosity into a chore.
The interface helps good ideas find their way back into your day.
Start with the smallest useful version
The prototype has only a few verbs: save, sort, annotate, and return. A saved link gets a title, a source, and an optional note. A handful of simple shelves keep “to read” separate from “keep close.” There is no score attached to any of it.
This narrowness is the point. Every extra feature carries a small tax: another decision, another notification, another surface to maintain. A calm tool is not empty; it is opinionated about what deserves to stay out.
Leave it a little unfinished
The best part of building locally is that the product can remain close to the problem. I can notice which moments feel awkward, change one thing, and use it again. No launch campaign is required. No dashboard needs to prove that reading is happening.
For now, the project is a modest experiment in keeping promises to your future self: save the thing, make it findable, and leave enough room for the next idea.