Interests

Interests are how you tell Scour what you care about. Add a few topics, and Scour finds relevant content across thousands of sources by combining two kinds of matching: what an article means and what it actually says.

What Interests Do

Unlike other platforms where clicking one article can flood your feed with a topic, Scour only shows content based on interests you've explicitly added. Your clicks and reactions influence what gets suggested, but only interests you choose to add affect your feed.

This keeps you in control. No more wondering why your feed suddenly changed after clicking on one random article.

Creating Interests

Start by adding 3-5 topics you genuinely want to read about. Interests can range from wide areas to narrow niches:

• Wide topics: "AI" will surface content about machine learning, LLMs, robotics, and more
• Narrow topics: "Rust async runtime internals" will find exactly that
• Personal topics: "Home Cooking" or "Photography" are great too. Scour isn't just for tech or work

Here are some examples of what other users are interested in:

Not sure what to add? Browse popular interests to see what others have found valuable. You can copy any interest with one click.

Feel free to add as many interests as you like. Some users have 500+. Scour automatically diversifies your feed so no single topic dominates.

Using Keywords

Each interest has an optional "keywords" field. Use it to refine or disambiguate:

• Disambiguation: "Rust" with "programming, cargo, crates" clarifies you mean the programming language rather than oxidation
• Narrowing scope: "Cooking" with "Italian, pasta, Mediterranean" focuses your results

Keywords influence both sides of matching: they shape the lexical filters and tilt scoring toward articles that actually use that vocabulary.

Tips for Good Interests

• Labels must be unique — You can't have two interests with the same label. Use distinct names like "AI Research" and "AI Tools" instead of both being "AI".
• Labels appear in your feed — When a post matches your interest, the label shows up below the title. Keep them concise.
• Acronyms work — Use "LLM" as a label with "large language models, GPT, Claude" as keywords.

How Matching Works

Scour combines two signals to compare your interests against articles:

• Semantic matching finds articles whose meaning is close to your interest, even when they use different words. "Machine Learning" matches articles about "neural networks"; "PyTorch" matches relevant content phrased differently.
• Lexical matching checks that an article actually uses your interest's vocabulary in the right sense, then weighs how much it's really about the topic. This is what makes "Rust" the programming language match programming articles, not articles about the Rust Belt or metal corrosion.

Both scores combine into one. How much each side weighs depends on the interest's specificity tier (see below).

Specificity Tiers

Each interest has a specificity setting that shifts the balance between semantic and lexical matching. You can change it by editing any interest.

• Broad — Semantic matching does most of the work. Good for wide topics like "Technology" or "Culture" where relevant articles often talk around the topic rather than naming it directly.
• Normal (default) — A balanced mix of semantic and lexical matching. Articles need to be on-topic by meaning and generally use the topic's vocabulary.
• Specific — Lexical matching does most of the work. Use this for precise terms like "Rust" or "PostgreSQL" where you want articles that explicitly discuss the topic with its actual name.

Scour picks a specificity when you create an interest. If results feel too noisy or off-topic, try Specific to require the topic's actual vocabulary. If you're missing content you'd expect, try Broad to let meaning carry more weight.

For the full picture of how content gets ranked, see How Ranking Works.

Interest Recommendations

As you use Scour (clicking articles, reacting to posts, subscribing to feeds) the system learns what topics might interest you. You'll see recommendations on your interests page and sometimes while browsing your feed.

Here's what a recommendation looks like:

Accept (+) to add it to your interests, or reject (×) to help Scour learn what you don't want.

Suggest Interests from Your Feeds

If you've subscribed to feeds or uploaded an OPML file, Scour can analyze your subscriptions to suggest interests automatically. This is a quick way to bootstrap your interests, especially if you're migrating from another reader.

Use the "Suggest from my feeds" link on the Interests page to get suggestions based on your current subscriptions. OPML uploads also trigger suggestions automatically.

Still have questions? Or feedback on these docs? Please let me know!