A EU-made Kagi alternative wants your money — and the thread wants to know who powers the search
Uruky, a privacy-focused European search engine pitched as a Kagi alternative, shipped image search and an account that's 'just an ID'. The Show HN turned into a clinic on what actually makes a search engine: not privacy, but results — and where those results come from.
The pitch, and the founder in the replies
Uruky is a European, privacy-first search engine explicitly framed as a Kagi alternative, and this Show HN announced image search plus a new URL feature. Its privacy hook is an account that's "just an ID" — no personal data attached. The founder participated directly, fielding UI feedback and explaining why certain features (like custom result ranking) are gated by what upstream providers allow.
The thread's reality check: privacy is not the product
The most bracing comment refused the premise that privacy alone wins:
"I live in the EU and I use kagi… I need a search engine to find stuff and kagi is better than google for the things I and my AI agents are searching for. If you don't get that no amount of better privacy is worth my time." — a Kagi user on what matters
A VC focused on European tech sovereignty sharpened it: founders conflate "EU-made alternative" (a real need) with maximalist privacy theater like accepting cash by mail, and end up building something that can't compete on search. That tension showed up at the checkout, too — one reader noted that an "anonymous ID" account stops being anonymous the second you pay with a personally-identifiable method.
The most consequential exchange was about provenance. A reader who boycotts Kagi over its use of Yandex asked Uruky for its sources, and got a concrete answer: Uruky federates Mojeek, Marginalia, EUSP (Ecosia/Qwant), Linkup and Serper for web, and Pixabay plus Serper for images — no Yandex. For the privacy-maximalist crowd, that's the differentiator that actually differentiates, more than the anonymous ID.
What this Show HN really exposed
Uruky's honest problem isn't trust — it's that search is a results game, and results come from indexes you either build or rent. The thread's verdict is a useful one for any "ethical alternative" startup: users will pay for sovereignty and privacy, but only after the engine clears the bar of finding the thing. Pick which upstreams you federate carefully, because in 2026 that choice — Yandex or not, EU indexes or not — is the product.
Get breakdowns like this every day
The top Hacker News stories, unpacked — by email, Telegram, or RSS.
An IPv6 address can smuggle a shell command, and a valid-IP check waves it through
Xe Iaso argues IPv6 zone IDs in URLs are a mistake because the '%' collides with URL escaping. The thread escalated it into a security lesson: a string the standard library certifies as a valid IPv6 address can carry 'whoami' straight into your shell.