Current Issue
This Month's Print Issue

Follow Fast Company

We’ll come to you.

2 minute read

This AI Librarian Organizes Your Bookmarks With Machine Learning

Stash is a great new service that automatically categorizes all the links you save.

  • 01 /06
  • 02 /06
  • 03 /06
  • 04 /06
  • 05 /06
  • 06 /06

The internet bookmark aspires to be the digital equivalent of the sticky note you paste on an important page in a college textbook. But that's where the metaphor falls apart. If the internet is a college textbook, it's an infinitely dense one, where every chapter covers a different subject, and someone has stuck a sticky note on every page.

The problem that nearly every bookmarking site inevitably needs to solve is one of organization. Pinterest does this with boards; other bookmarking services, like Pinboard, do so with tags. All of these solutions, though, require users to rigorously categorize their own bookmarks, which when you think about it, is kind of the equivalent of asking someone to bind and title every physical book in their library before they actually put a bookmark in it.

Stash is a new bookmarking site that has an innovative solution to the whole organization problem. Stash gives your bookmarks its own AI librarian, which uses machine learning to automatically categorize all of your bookmarks in different categories: for example, articles, or recipes, or products, and so on. From there, if you want to further classify your bookmarks, you can add tags, but even if you just follow a "throw everything into a bucket" school of bookmarking, Stash will make it easy for you to browse your bookmarks by category, as well as search for them later.

Founded by Rahul Shah and Christopher Goes out of Ithaca, New York, two years ago, Stash wasn't always a bookmark manager. It actually started as an analytics engine for the news and media industry. "Our goal was to help publishers qualitatively understand their audience by analyzing the meaning of the content they were reading," says Shah. "In the process, we had built some pretty solid technology for parsing and understanding the meaning of webpages." About eight months ago, they realized that the AI Librarian they had built was better applied to consumer tech, "to help us keep track of the webpages that we were finding on a daily basis—which were a complete mess."

Besides its AI Librarian (LibrAIrian?), Stash has some other handy features. A social component allows you to send bookmarks to other people. Bookmarks sent to you, meanwhile, go into an inbox, where they can be accepted or deleted. Accept an incoming bookmark, and you can even search for it by the person who sent it to you, useful for when your significant other asks if you read that article she sent you a few months ago yet. Stash even borrows some tricks from email apps like Mailbox and Inbox by Gmail to give you the option to "snooze" bookmarks. Don't have time to read something now? Stash will remind you about it in another week.

Stash isn't perfect. Contrary to what some hyperbolic headlines have claimed, it doesn't streamline formatting to give articles a better reading experience, so if you do a lot of internet reading, you'll still probably rely on services like Pocket or Instapaper. But the mobile app is surprisingly great, allowing you to easily add any link to your Stash through your device's share menu. If you juggle a lot of bookmarks, sign up for Stash's waiting list here. It might just become your new default bookmarking site.

The Fast Company Innovation Festival