Update, October 2026: Tempmarks is closed. After ten months it had a handful of kind users and no sign of more, so I shut it down; tempmarks.com now shows a short goodbye. The post below is how it looked in January, with a note at the end on what I took away.
Somewhere in my browser there's a folder called read later. It has years of links in it. I couldn't tell you what's inside, because I never open it; I only add to it. Every bookmark in there felt important for about a day, and then it became sediment.
That's the flaw in every bookmark manager ever built: they all assume a saved link is worth keeping forever. It isn't. Information has a half-life. If a link matters, you'll come back for it this week. If you don't, it was never important, and now it's clutter with a favicon.
So for January's project I built Tempmarks, a bookmark manager where links expired by default.
How it works
Every saved link lives for seven days. It stays visible and one click away the whole time, and if you haven't done anything with it by the end of the week, it quietly disappears. No cleanup sessions, no guilt, no archive growing in the dark. You can save full links or short notes, and if something earns its keep, you promote it to permanent storage with one click.
That's the entire model. Temporary by default, permanent only if earned.
Seven days turned out to be the right constraint. It's long enough to actually get to the thing and short enough that you can't hoard. Instead of "should I keep this forever?", the only question a bookmark ever asks you is "is this worth my attention this week?" That's a much easier question, and most links fail it.
Where it fits
I don't use Tempmarks instead of Notion or Obsidian. I use it in front of them. It's the inbox where links prove themselves before they're allowed anywhere near long-term storage: whatever survives a week of scrutiny gets promoted and filed properly, and everything else expires without me ever making a decision about it.
If you spend your days saving documentation, threads, and RFCs you swear you'll get back to, that filter does a surprising amount of work.
What shipped
A web app that worked on desktop and mobile, a Chrome extension, OAuth sign-in, a free tier and a paid one nobody bought. No ads, no selling data, minimal analytics. Behind it a NestJS API on App Runner, Postgres on RDS, a Lambda that expired links on the hour, and a marketing site with eight blog posts arguing that your bookmarks don't work.
What happened
Not much, which is the whole story. Ten months in there were a handful of users, no paying ones, and no curve to point at. I still believe the seven-day constraint is right. I no longer believe enough people feel the problem strongly enough to install something for it; a folder called read later is annoying, but it's free and it never asks anything of you.
The filter itself survives. I still give a link a week to prove it matters before it's allowed anywhere near long-term storage; it just doesn't need a product to do that.
So long
The app, the API and the extension are gone. Nothing of yours was sold and nothing is kept running. If you used it, thank you, and if you want to talk about it, write to support@tempmarks.com.