QWE AI Academy Why do I keep clicking search results I can’t read?
You’ve felt it. Query something timely, get a perfect-looking headline from a major outlet, click… and hit a subscription wall. Time wasted. Tab closed. Repeat.
Kagi shipped a direct fix on August 21, 2026: remove paywalled links from search results automatically. Control label on Search settings: Exclude paywalled websites. The auto-exclude feedback thread had been asking for a global filter instead of endless manual blocking.
Not a changelog paraphrase. Below: flip the switch, keep the outlets you actually pay for, run one messy query, and know the three limits the release note leaves open.
Why the old workarounds still fall short
Kagi already beat generic search on this problem. A paywall indicator showed up next to many locked domains years earlier (changelog history ~Dec 2023, later paywall signals on /news too). Shield icon on any result → Block or Lower forever. Quick actions still include Open page in Web Archive (Wayback) and Continue in Assistant. On the domain-ranking tips post, users said they blocked every paywalled domain they couldn’t access – and ranked that control as a top reason they pay for Kagi.
Those tools still work. They’re slow when the SERP is half metered walls and half hard walls. Manual blocking is whack-a-mole; new vertical sites show up weekly. The indicator saves a click, but the noise stays in the list. Archive helps after you already cared about a URL. It doesn’t clean the result set.
One switch at the list level. Fewer dead ends before you click. That’s what changed.
Turn on Kagi’s setting for removing paywalled links
Active account required – trial is enough. Plans as of late August 2026: Trial (100 searches), Starter $5/mo for 300 searches + 300 AI interactions, Professional $10/mo unlimited searches + standard Assistant, Ultimate $25/mo with premium Assistant models (plus tax). Fair Pricing puts a full credit on fully unused months. Numbers live in the plan types docs.
- Sign in at kagi.com.
- Open Settings (profile / hamburger).
- Search settings: kagi.com/settings/search.
- Find Exclude paywalled websites (same wording as the changelog anchor).
- Enable it. Next searches pick it up – no restart.
Pick a query you usually lose to walls: breaking news, niche industry analysis, long-form explainers. Glance at free-to-read density before vs after. Whole loop takes a minute.
Pro tip: the exclude is global. Right after you enable it, jump to the next section and Pin/Higher the domains you subscribe to before you judge the SERP.
Keep the walls you actually pay for
The catch: a blunt filter also hides The Athletic if you have a sub, or the trade journal your company buys.
Shield icon on any result → Higher or Pin for sites you can read. Full list later under Settings → Personalized Results (user_ranked). Official options, per the Website Info & Personalized Results docs: Block, Lower, Normal, Higher, Pin. Community feedback wanted a pre-approved exception list for paywalled sites the user can access – this older ranking system is that list.
Real search: industry news without the subscription tax
Chip shortage. Regulatory filing. Without the filter, the top often looks like three major outlets you don’t pay for, two aggregator rewrites, one free primary source buried lower.
Exclude on. Known-good free domains (gov pages, company IR, solid independents) left at Normal or Higher. Free primary and secondary sources climb. Fewer dead-end tabs. When a locked piece still looks essential, the result quick actions still offer Open page in Web Archive or Continue in Assistant – same actions the personalized-results docs list.
Honest limits and how to live with them
Powerful. Not magic.
- Detection is a black box for now. Changelog wording is only “removing paywalled links from search results automatically” with the Exclude label. No public domain list, no confidence scores, no statement on whether it uses a static list,
isAccessibleForFree-style structured data, crawl signals, or a mix. Soft walls, metered freemium blogs, login gates, and hard subscription walls may not behave the same. First week = calibration. Freemium blog you like vanished? Raise it manually. - Other surfaces. The note sits under Kagi Search. Whether News, lenses, Assistant web sources, or image results fully honor the same flag is not stated. Older items separately covered paywall signaling on /news and blocked-domain leakage into Quick Answer – so don’t assume parity until you check the surface you actually use.
- Coverage trade-off. Topics owned by subscription journalism get thinner. Expect more manual overrides on those queries, not a perfect free-only encyclopedia.
Cleaner default SERP. Cost: whitelist the walls you pay for, and accept that a useful locked source stays hidden until you do.
FAQ
Does the exclude paywalled setting work on the free trial?
Yes. Trial gets the full settings surface for the 100 allotted searches. Turn it on while you evaluate.
I subscribe to two paywalled sites – will they vanish?
They can. Classifier flags the domain, no override yet, and the hit disappears from the list.
Concrete fix: enable Exclude, search a known article from sub #1, shield → Higher or Pin. Repeat for sub #2. You’re done – ranking rides on top of the global filter.
Is this better than just using an archive extension or uBlacklist?
Wrong comparison if you treat them as substitutes. Extensions act after the SERP loads or on the destination page. Kagi drops the links from the result set, so you never rank or click them in the first place.
Extensions still win for one-off locked URLs that slip through and for path-level rules domain ranking doesn’t express. Most people who already live in blocklists will keep both: Kagi for the default clean list, extension for surgical leftovers.
Enable Exclude paywalled websites on Search settings, Pin the two outlets you actually pay for, then rerun one messy news query you abandoned last week. Thirty seconds. You’ll know if it earns a permanent slot in your defaults.