Follow a sector, but only in publications you trust

/stories in hybrid mode. The parameters matter less than the reasoning behind them, so that is most of what is below.

The request

HTTP

GET /stories
  ?q=<sector terms>
  &mode=hybrid
  &source=<domain>
  &from=<now - 3d>

# repeat per outlet, merge client-side

Why these parameters

The outlet filter is the most under-used control in the API. Restricting to a known publication set is usually a better proxy for quality, region or editorial position than any other field, and it turns an unbounded topic query into something you can actually reason about.

The obvious approach, and why it loses

The obvious approach is to filter on quality indirectly — by topic, by signal range, by tone — and hope it correlates with the publications you would have chosen. It correlates weakly. Editorial standards are a property of a newsroom, not of an article's vocabulary, and every proxy for them is downstream of something else: signal ranges track subject matter, tone tracks register, topic tracks classification. Naming the publications puts the judgement where it belongs and makes it inspectable, which matters because the judgement is the part a colleague will want to argue with.

When it returns too much, or too little

Too little is the expected state and usually correct: a constrained query over five outlets returns what those five published, and widening the window is the right response rather than adding outlets in the moment. Too much means one publication in the set is a high-volume aggregator; check the per-outlet counts before assuming the sector got busy. The real tuning here is the set itself, and it needs a scheduled review rather than a parameter. Twice a year, run the same query unconstrained, read what the constraint excluded, and decide deliberately whether each exclusion is still one you would make.

How you would know it is working

Measure what the constraint costs you, not what it gives you. Once a quarter, run the query with and without the outlet filter over the same window and read the difference: those are the stories your trusted set did not carry. Some will be noise, which is the filter working. Some will be original reporting from a publication nobody on the team had heard of, which is the filter costing you the thing it was supposed to protect. A set that excludes nothing interesting for two reviews running is probably too wide; one that excludes something important every time is too narrow, and neither shows up in the feed itself.

When this is the wrong tool

An outlet set is a decision you have to revisit, and most people never do. Publications are acquired, editorial teams turn over and titles fold, so a set curated two years ago is quietly measuring a different thing than it was — and because it keeps returning results, nothing tells you. This recipe is also the wrong tool if what you need is discovery: constraining to publications you already trust guarantees you will never find the outlet that covers your sector best and that you have not heard of. Run it alongside an unconstrained query periodically and compare what the constraint excluded, rather than treating the set as settled.

Running it for real

Response fields, failure modes, adaptations and scheduling cost: Follow a sector, but only in publications you trust, in the docs.