Kagi Will Delete Paywalled Articles From Your Search Results. Should You Let It?

You search for something. You click the first result. Subscribe to keep reading. You click the second. Subscribe to keep reading. Kagi, the paid search engine, has a blunt answer to this: just remove those sites from your results altogether. It is a small feature that opens a surprisingly large trapdoor.

One thing up front. There is almost no recent community discussion to draw on here — no substantial Reddit threads in the last month, no usable X data. So this is less a read of live sentiment than an attempt to take the argument apart and see how it’s built.

What Kagi Actually Does

Kagi is a subscription search engine with no ads. You are the customer, not the advertiser. That inversion lets it do things Google structurally cannot, and the most interesting one is per-site ranking control.

You can Raise a domain, Lower it, or Block it outright. Layered on top of that is a single option that handles every paywalled site at once. One checkbox, and the New York Times, the Wall Street Journal, and Bloomberg quietly stop appearing.

Technically this is nothing — a domain list and a filter. But handing users a switch that says make this category of information invisible to me is a different kind of product decision than nudging a ranking up or down. It’s an editorial power, delegated.

The User’s Case Is Hard to Argue With

The pro-filter argument is tight, and it rests on one thing: time.

Search exists to answer questions. A link you can click but cannot read isn’t an answer — it’s closer to an ad for a subscription. If 4 of your 10 results are walled off, the real quality of that result page is 60%. When you’re paying a monthly fee for search, asking the engine to drop the ones you can’t open is not an unreasonable demand.

There’s a sharper edge to this too. Plenty of publishers serve the full article to Googlebot and a paywall to the human who arrives from the same search. They want the ranking that full-text indexing buys and the revenue that blocking readers buys. This is cloaking, and it’s explicitly against search engine guidelines — news organizations have simply enjoyed a carve-out for years. Against that backdrop, fine, we won’t show your link at all is a fairly coherent response.

The Other Side Isn’t Weak Either

The problem is that the filter doesn’t only remove annoying links.

Outlets with paywalls are, broadly, the outlets that spend money on reporting. Foreign bureaus. Six-month investigations. Lawyers on retainer for when the subject of the investigation calls their lawyers. Meanwhile, a large share of the paywall-free web is in the business of summarizing and re-circulating that work. Filter out the paywalls and you don’t remove the story — you remove the original and keep the copy.

For any individual user this isn’t a loss. You still learn what happened. But the click goes to the aggregator. So does the subscription conversion that never happens, and the brand impression that never lands. The outlet that paid for the reporting gets nothing; the one that rewrote it gets everything. This is the tragedy of the commons, in its most textbook form: every individual choice is rational and the aggregate outcome is ruin.

And there’s a second-order problem. Once the filter is on, you can’t see what you’re not seeing. If the best piece on the topic was result number three, its absence doesn’t register as absence. You just get a page with 8 results and no reason to suspect anything is missing. Invisible and nonexistent look identical from the inside.

One Genuinely Strong Rebuttal

The filter is off by default. You have to turn it on. It applies to your account and nobody else’s. That is categorically different from Google algorithmically demoting a publisher for everyone at once.

The closest analogy is the ad blocker. Ad blocking was attacked on exactly these grounds — you’re breaking the business model that funds the content — and the criticism was accurate. It did break it. But the consensus settled anyway: most people now treat blocking as a legitimate user choice, because most people concluded the fault lay with sites that stacked 40 trackers and a full-screen interstitial on top of a 600-word article, not with the person who declined to load them.

The paywall filter borrows that argument wholesale. The cloaking came first; the filter is the reaction. It’s a real point. But ads and reporting are not the same input. Strip out the ads and the article still exists. Strip out the reporting budget and there is no article to block.

The Real Problem May Be Upstream

Step back and this argument is really pointing at something else: the structural ceiling on subscriptions as a model.

Nobody subscribes to ten news outlets. Nobody pays $15 a month to read one article one time. Which means that in practice, most paywalls function as a sign reading do not read this article — the filter just takes the sign down before you walk into it.

Per-article micropayments, cross-publisher bundles, public funding: all of it has been debated for two decades and none of it has stuck. Until something fills that gap, users will keep having good reasons to turn the filter on and publishers will keep having good reasons to build the wall higher. Both sides are behaving sensibly. The outcome gets worse anyway.

Where I Land

Kagi’s paywall filter isn’t a bad feature. If anything it’s an honest one — a piece of UI that accurately reports how broken news distribution has become. Things have to be pretty far gone before a search engine needs a setting for hide the links you can’t read.

But every time you flip that switch, you cast a small daily vote for the outlets that don’t pay for reporting. Clean results where every link opens, or messier results where at least you know the good article exists. I haven’t decided which one I want. Have you?