Search for a list of AT relays, and Google plays innocent and vanishes

July 12, 2026, 03:09 — Source: "I'm a Little Painter 2" (NetEase self-media account), Beijing Translated from Chinese. Original: https://www.163.com/ (NetEase)

The same search query — "list of public atproto relays" — instantly surfaces five professional entry points on DuckDuckGo, while Google coughs up only three irrelevant links. And the one site that did make the rankings? Open it up and — sorry, it doesn't have a list either. This isn't a one-off, either; it's been a black hole for this basic query for quite a while.

A user brought this up on Hacker News, and it instantly hit a nerve in tech circles. The AT Protocol (Authenticated Transfer Protocol), a new pillar of decentralized social networking, has developers rolling out new sites and tools in its ecosystem every week — yet the biggest search engine acts as if none of it exists. The irony is off the charts: a protocol bound tighter than any to the open web is being completely walled off at the information entry point.

But before you rush to accuse Google of evil, netizens have already dug through several possible explanations, each with its own sharp streak of snark.

① What does Google actually give you? Three bounced checks

DuckDuckGo's results are clean and clear: firehose.directory, atproto.at/relays, atproto.wiki, pulsar.feeds.blue, leaflet.pub — all public lists of AT protocol relays or directly related documentation. On Google's side, only three results appear on the search page, and the only one directly related to AT, firehose.directory, has no relay list to be found once you click through. The other two are either about relays for other protocols or completely unrelated. In other words, it's not that Google hasn't crawled the content — it crawled it and then buried the genuinely useful stuff below the waterline where no one can see it.

② Is personalization the scapegoat? It fails in incognito mode too

Some people immediately blamed "personalized search" and proposed testing in incognito mode. The result: two people reproduced exactly the same missing results in private windows — Google still wouldn't show those specialized sites. Someone else said they got 2,300 results, but clicking through revealed nothing but Tor relays, Nostr relays, IPFS relays, and fediverse relays — everything except a list of AT protocol relays. Personalization can certainly muddy the waters, but it's clearly not the main culprit here; this looks more like a systemic ranking problem.

③ Did AI Overviews intercept it? Maybe Google would rather be the answer itself

A key clue surfaced: some users found that the AI Overview for the query did include generated content listing relays — while the organic search area remained blank. This invites the suspicion that Google, in tuning its algorithms, is increasingly inclined to let its own "list generation" feature answer questions directly, drastically demoting external sites with the same kind of content. You want a list? The AI will cobble one together for you — as for those carefully maintained specialist pages, sorry, no room. This "produce it yourself, consume it yourself" closed loop is death by a thousand cuts for small sites that depend on search engine traffic.

④ 2,300 results, and not a single useful one

The "I got 2,300 results" comment looks like a rebuttal at first glance, but on closer inspection it's even more damning. Those 2,300 results are stuffed with Tor, Nostr, IPFS, and other relay links — you can scroll to page six without finding a single AT protocol relay list. The number 2,300 is itself dark comedy: Google can bury you in mountains of information fragments while precisely kicking the most direct, most relevant handful of pages out of view. This isn't "can't find it" — this is deliberate avoidance.

⑤ The great irony of open protocols: the content exists, the entrance is dead

The AT Protocol's championed ideals — decentralization, data ownership, the open web — are attracting waves of developers, with new sites launching every week. But the weakest link in this whole ecosystem is precisely: how does the outside world find it? If Google keeps failing to index this content, ordinary users have almost no way of reaching the latest developments in the decentralized world through search. A movement founded on openness ends up with its throat squeezed by a closed search algorithm — even originally neutral observers might start to wonder: is this a technical flaw, or some kind of intentional downranking?

At the end of the day, a search engine shouldn't be a black box accountable only to itself. When a query explicitly asks for a "list" and a whole crowd of public resources collectively goes invisible, what we need isn't just a bug fix — it's an explanation.