Three primary search engines index .onion sites on Tor:
Ahmia (ahmia.fi) is operated by Finnish security researcher Juha Nurmi with partial Tor Project funding. It actively filters abusive content and offers a clearnet interface at ahmia.fi where you can search .onion sites before connecting to Tor. Best for moderated, curated results.
Torch (xmh57jrknzkhv6y3ls3ubitzfqnkrwxhopf5aygthi7d6rplyvk3noyd.onion) has operated since 2013, making it one of the oldest Tor search engines. It claims to index 1-3.2 million .onion pages with minimal content filtering. Best for comprehensive, uncurated coverage.
Haystak claims the largest index at over 1.5 billion pages. It offers a free basic search tier and a premium tier with advanced search operators, cached page views, and historical snapshots. No confirmed .onion address is available from provided sources.
Ahmia: Curated Search with Clearnet Access
Ahmia is an open-source dark web search engine operated by Finnish security researcher Juha Nurmi. It has received partial funding from the Tor Project itself, giving it credibility within the privacy community.
Ahmia actively filters abusive content from its index. This moderation removes child abuse material, scams, and other harmful content before it appears in search results. You get a cleaner, safer search experience compared to unfiltered engines.
The clearnet interface at ahmia.fi allows you to search for .onion sites before connecting to Tor. You can preview search results, verify site descriptions, and decide whether to proceed—all from your regular browser. Once you find a site you want to visit, copy the .onion address and open it in Tor Browser.
This two-step approach reduces exposure to malicious content. You avoid clicking unknown links directly in Tor Browser, where malware and phishing sites are common. Instead, you research first, then visit only the sites you've vetted.
Ahmia is best for users prioritizing moderated results. If you want to avoid illegal content and prefer a curated index, Ahmia offers the most filtered search experience among major dark web engines. Researchers, journalists, and privacy advocates often start with Ahmia for this reason.
The trade-off is coverage. Ahmia's filtering removes not just harmful content but also controversial or borderline material. You won't find everything the dark web hosts, only what passes Ahmia's moderation standards.
Torch: Oldest Engine with Minimal Filtering
Torch is one of the oldest Tor search engines, operational since 2013. Its longevity makes it a known quantity in the dark web community. Many users trust Torch because it has survived while other engines disappeared.
Torch claims to index over one million .onion pages. An alternate source cites around 3.2 million pages. The discrepancy reflects the difficulty of measuring dark web content, as sites appear and vanish constantly. Either figure represents substantial coverage compared to smaller engines.
Torch performs minimal content filtering. Search results include the full spectrum of hidden services without editorial curation. You see everything Torch has indexed, regardless of legality or content type. This unfiltered approach appeals to users who want comprehensive coverage and are willing to navigate risks themselves.
The .onion address for Torch is xmh57jrknzkhv6y3ls3ubitzfqnkrwxhopf5aygthi7d6rplyvk3noyd.onion. You must access this address through Tor Browser. Copy the address exactly; typos or fake versions can lead to phishing sites.
Trade-offs of minimal filtering:
Comprehensive coverage: Torch indexes sites that Ahmia and other moderated engines exclude. You find obscure forums, niche services, and controversial content.
Unfiltered results: You encounter illegal marketplaces, scam sites, and potentially harmful content. Torch does not protect you from these risks.
User responsibility: You must verify sites independently. Check .onion addresses through multiple sources, avoid downloading files, and never enter personal information without confirming legitimacy.
Torch is best for users who need broad coverage and understand the risks. If you're researching dark web ecosystems, tracking specific services, or exploring beyond curated indexes, Torch provides access other engines don't. However, you must exercise caution with every search result.
Haystak: Largest Index with Premium Features
Haystak claims to maintain the largest index of .onion sites with over 1.5 billion pages indexed. This figure far exceeds Torch's claimed 1-3.2 million pages, positioning Haystak as the most comprehensive dark web search engine by raw numbers.
Haystak offers both a free tier and a premium tier. The free tier provides basic search functionality similar to other engines. You enter a query, receive a list of .onion addresses, and visit sites through Tor Browser.
The premium tier adds advanced search operators, cached page views, and access to historical snapshots of .onion sites. These features serve specific use cases that free search engines don't address.
Advanced search operators let you refine queries with Boolean logic, exact phrases, and exclusion terms. If you're researching a specific topic across multiple sites, operators help you filter noise and find relevant content faster.
Cached page views store copies of indexed pages on Haystak's servers. If a site goes offline, you can still view the cached version. This is useful when tracking ephemeral services or verifying information from sites that frequently change addresses.
Historical snapshots archive .onion sites over time. You can see how a site's content evolved, verify past claims, or access information that has since been removed. Researchers studying dark web trends, law enforcement tracking criminal activity, and journalists investigating stories all benefit from historical data.
Use cases for premium features:
Academic research: Scholars studying dark web communities, marketplaces, or communication patterns need historical data to analyze trends over time.
Investigative journalism: Reporters verifying sources or tracking illegal activity benefit from cached pages and historical snapshots when sites disappear.
Security analysis: Cybersecurity professionals monitoring threat actors or tracking malware distribution use advanced operators to filter large datasets.
Legal compliance: Organizations ensuring they don't inadvertently link to illegal content can verify site history before referencing .onion addresses.
Haystak's premium tier is not free, but the sources do not document current pricing. If you need these features, visit Haystak's .onion address through Tor Browser to check subscription options.
Haystak is best for users who need the largest possible index and are willing to pay for advanced tools. If your work requires historical data, cached access, or sophisticated search queries, Haystak offers capabilities other engines lack. For casual browsing or basic searches, the free tier or other engines may suffice.
Understanding Dark Web Scale and Search Limitations
Tor Metrics provides the most reliable data on dark web scale. As of January 1, 2025, there were 787,890 unique v3 .onion addresses in the network. By May 2025, this number rose to 863,711. On a typical day, roughly 700,000 to 900,000 unique v3 .onion addresses are visible.
These numbers come from aggregated statistics reported by Tor relays acting as onion-service directories. The Tor Project extrapolates the total count if at least 1% of relays report data. This methodology provides a consistent estimate, but it counts addresses, not active websites.
The reality behind the numbers:
Estimates of genuinely active dark web sites at any one moment run into the low thousands, not the hundreds of thousands. Most unique addresses are short-lived, duplicated, or belong to services that were never public.
Why the discrepancy matters:
Temporary addresses: Users generate .onion addresses for one-time file sharing, private communication, or testing. These addresses appear in Tor Metrics but never host public content.
Mirrors and backups: A single service may operate multiple .onion addresses for redundancy. If a site has five mirrors, Tor Metrics counts five addresses, but it's one service.
Abandoned sites: Operators shut down services but leave addresses in the network until they expire. These addresses remain visible in metrics but lead to offline sites.
Development and testing: Developers run .onion services locally or on private networks. These addresses are visible to the Tor network but not accessible to the public.
Why no search engine can be comprehensive:
Dark web search engines face challenges that clearnet engines don't:
No central registry: There is no authoritative list of .onion sites. Search engines must discover sites through crawling, user submissions, or monitoring Tor traffic.
Ephemeral content: Sites appear and disappear within hours or days. By the time a search engine indexes a site, it may already be offline.
Access restrictions: Many sites require authentication, invitations, or specific entry points. Search engines can't index content they can't access.
Operator resistance: Dark web operators often block crawlers or use robots.txt to prevent indexing. They value privacy and don't want their content discoverable through search.
Resource constraints: Crawling the dark web is slow and resource-intensive. Tor's onion routing adds latency, and many sites are unreliable. Search engines can't index everything even if they wanted to.
Why results vary dramatically between engines:
Each search engine uses different crawling strategies, filtering policies, and update schedules:
Ahmia filters abusive content, reducing its index size but improving result quality.
Torch indexes broadly with minimal filtering, capturing more sites but including illegal and harmful content.
Haystak claims the largest index but may include many inactive or duplicate addresses.
A site indexed by Torch may not appear in Ahmia if it violates moderation policies. A site indexed by Haystak may be offline when you try to visit it. No single engine provides a complete, current view of the dark web.
Practical implications:
When searching for .onion sites, use multiple engines. If you don't find what you need in Ahmia, try Torch or Haystak. Cross-reference results to verify addresses and check whether sites are still active.
Understand that search engines show only a fraction of the dark web. Many services are invite-only, shared through private channels, or intentionally hidden from search. If you're looking for a specific service, directories and forums may be more reliable than search engines.
The dark web is smaller and more fragmented than popular media suggests. It's not a vast hidden internet but a niche ecosystem with a few thousand active sites at any given time. Search engines help you navigate this ecosystem, but they can't index everything.
Q&A
- What are the best search engines for the dark web?
Ahmia is an open-source search engine operated by Finnish security researcher Juha Nurmi with partial Tor Project funding, offering moderated results and a clearnet interface at ahmia.fi. Torch, operational since 2013, claims to index 1-3.2 million .onion pages with minimal filtering, while Haystak claims the largest index at over 1.5 billion pages with free and premium tiers.
- How big is the dark web?
Tor Metrics recorded 787,890 unique v3 .onion addresses on January 1, 2025, and 863,711 in May 2025, with typical daily ranges of 700,000-900,000 visible addresses. However, estimates of genuinely active sites run into the low thousands, as most addresses are short-lived, duplicated, or never-public services.
Where to look next
These services are useful starting points for further research. Tools & services.