How TinyGem.org's Recommender Works with the Links You Save

Updated

TinyGem.org turns the links you bookmark into a source of recommendations: it automatically uses the links you save to surface other related content from manually curated sources. In other words, saving a link is the input, and a set of related reads is the output — you don't configure topics or fill out a profile to get suggestions. This suits people who already read places like Hacker News, keep a selective news diet, and want a way to keep track of pages they fear losing.

The mechanism: save first, recommendations follow

The site describes itself as a bookmarking service with a built-in recommender. The key relationship is direct:

  • Input: the links you save to your account.
  • Action: TinyGem automatically uses those saved links to surface related content.
  • Output: recommendations drawn from manually curated sources.

Because the trigger is your own saved links, the quality of recommendations depends on what you bookmark. Saving a handful of pages on one subject gives the recommender a narrow signal; saving across several interests gives it a broader one. The site does not describe a separate "training" step — the act of saving is what drives it.

Why "manually curated sources" matters

The recommendations aren't pulled from the open web at random. TinyGem states that related content comes from manually curated sources. That is the main trust signal on the page: a human has already decided which sources are worth drawing from, rather than an unfiltered crawl. It doesn't guarantee every recommendation will fit your taste, but it tells you the pool is bounded and intentional.

What you actually get

The homepage shows a "Recently Saved Links" feed — real examples include items from parool.nl, wired.com, a16z.news, arxiv.org, wsj.com, foreignaffairs.com, and others, each with a timestamp, the saving user, and a comment count. This gives you a sense of the reading diet the service is built around: news, policy, tech, and long-form analysis.

The page also offers a recommender you can try on any topic before committing. That's the practical way to test whether the curated sources match your interests: pick a subject you know well, run it, and judge the results.

How to use it, step by step

  1. Register an account. The site's navigation includes Login and Register; saving links is tied to a user account (the feed shows items "by" named users).
  2. Save links you want to keep. The stated purpose is keeping pages "saved forever" so you don't worry about finding the same material again.
  3. Save up to 10 links to unlock the personalized feed. The site says: "Register and save 10 links to get personalized curated feed from our AI." Until you hit that threshold, expect the general recently-saved feed rather than your own.
  4. Check the recommendations that follow. Once your saved links are in, the recommender surfaces related content from the curated sources.
  5. Verify fit with the topic tester. If you're unsure, use the "recommender on any topic" feature first — it's the lowest-commitment way to see the output style.

Common sticking points

  • Saving fewer than 10 links: the personalized feed is gated behind that count, so early results may feel generic.
  • Narrow bookmarking: if all your saved links are on one topic, expect recommendations clustered there — that's the design, not a fault.
  • Expecting a search engine: this is a recommender over curated sources, not an open-web search. Results are bounded by what the curators include.

Is it for you?

Choose TinyGem if you already bookmark reading material, want those bookmarks to persist, and would rather receive related reads from a curated pool than hunt for them yourself. Skip it if you want open-web discovery, or if you're unwilling to save at least 10 links to see the personalized output. The site does not publish pricing details on the page reviewed, so check the Register flow for any account terms before relying on it long-term.