22 ottobre 2024 rete tor 3 min di lettura

Motori di ricerca e siti onion: come funziona davvero la scoperta

Type an .onion address into Google and you get nothing. Not a 404, not a "no results" - just silence. The clearnet web's most powerful discovery machines are blind to an entire parallel ecosystem, and that blindness shapes how people actually find things on tor.

Why the big engines stay out

Indexing an onion site requires crawling it, which means running a tor client at scale, which means your crawler's traffic is indistinguishable from any other visitor. Google could technically do it. It chooses not to: the legal risk, the abuse surface and the near-zero advertising value make onion indexing a pure liability for a company that monetizes search. So discovery falls to a different breed of engine. Ahmia, a project born from a Finnish researcher's master's thesis, crawls onion services and - unusually - exposes its index on the clearnet too. Torch, one of the oldest, has been crawling since the Silk Road era from a server that has outlived most of what it indexes.

What crawlers can actually reach

Onion crawling is hard in ways clearnet crawling is not. There are no backlink graphs worth trusting, no sitemaps, no webmaster tools. A crawler follows links from directories and link lists, times out constantly, and indexes a lot of dead addresses. The coverage of even the best onion search engine is a fraction of what the name "search engine" suggests. Worse, the indexes age badly. Services vanish overnight, and a result set is often a museum. Research on onion crawlers has repeatedly found that a large share of indexed addresses no longer respond - which is exactly why a live status check belongs next to every search result.

The phishing layer on top

Onion search results are also a favorite hunting ground for scammers. Fake mirrors of well-known services buy or hack their way into indexes and link lists, then harvest logins from anyone who arrives trusting the search engine instead of verifying the address. The engine is a discovery tool, never an endorsement.

How to treat search results

The practical rules are boring but effective: treat every result as unverified until the address is confirmed against a primary source, prefer directories that PGP-sign their listings, and never enter credentials into a site you reached through a search engine for the first time. Our address verification guide covers the full checklist. Discovery on tor is a patchwork of volunteers, half-maintained crawlers and community link lists. It works - but it works like a bulletin board, not an oracle. The difference matters every time money is involved.

L’UE ripristina il Chat Control 1.0: la scansione dei messaggi torna nonostante la maggioranza abbia votato contro

Il Parlamento europeo non è riuscito a bloccare la ripresa del regime temporaneo e volontario di scansione dei messaggi il 9 luglio 2026 - 314 deputati hanno votato contro, ma la soglia della maggioranza assoluta non è mai stata raggiunta. Con i negoziati sul CSAR permanente che riprendono a settembre, la battaglia tra scansione e crittografia decide direttamente il futuro degli strumenti di anonimato.

Telegram vuole .gram: cosa succede quando l’infrastruttura criminale lascia le onion per un DNS controllato dalla piattaforma

Telegram ha chiesto all’ICANN il dominio di primo livello .gram, trasformando i nomi utente in domini risolvibili con siti generati da IA dietro. Poche settimane dopo la sospensione di t.me sotto pressione delle sanzioni, la mossa mostra come l’infrastruttura si allontana dagli onion service verso un DNS che i regolatori possono davvero toccare.

Kimwolf v7 e l’onion cablata nel codice: come i botnet hanno smesso di temere i sequestri

Il botnet Kimwolf v7 resuscitato risolve i suoi server di comando tramite Ethereum Name Service su cinque endpoint mescolati, e ricade su un hidden service Tor fisso se tutti falliscono. Un caso di studio sul perché gli onion service restano il canale C2 più resistente ai sequestri in mano ai criminali.