OnionDark Web Search Engines: Full Comparison 2026
On the regular web, finding information means typing a few words into Google and scanning the top results. On the dark web, it is not that simple. Google does not index .onion sites, and the specialized search engines that exist for the Tor network each have their own policies, strengths, and weaknesses. This article provides an in-depth comparison of the main dark web search engines: Ahmia, Torch, Haystak, DuckDuckGo on .onion, and a few lesser-known alternatives. For each engine, we cover its history, catalogue, filtering policy, concrete advantages, and limitations, so you can choose the right tool for your needs.
⚫ This page is the showcase. The rest is elsewhere. Tor access →Why specialized search engines?
Before comparing engines, let us recap why the dark web needs its own search tools. Three technical reasons make it impossible for Google to index .onion sites.
The Tor protocol. Googlebot, Google's indexing crawler, browses the web via standard HTTP and HTTPS protocols. It does not support the Tor protocol, which is required to access .onion services. Modifying Googlebot to support Tor would require considerable investment for virtually no commercial return.
The absence of public DNS. Standard sites are discovered by Google via DNS records, sitemaps, and inbound links. .onion addresses are not listed in any DNS — the only way to discover them is to follow links from other .onion sites or from public directories. That requires running Tor oneself and crawling the network, which Googlebot does not do.
Economic incompatibility. Google's business model rests on targeted advertising, which requires profiling users. An anonymous environment like Tor is structurally incompatible with that model. For a detailed description of the difference from the regular web, see our article Dark Web vs Deep Web.
Specialized engines have therefore emerged, each with a different approach. Each is itself hosted on .onion, each has built its own Tor crawler, and each applies its own filtering policy. These are the four main engines we cover below.
Ahmia: the ethical reference engine
Ahmia is the most widely cited .onion search engine for users who want an ethical approach. Created in 2014 by Finnish researcher Juha Nurmi at the University of Tampere, Ahmia grew out of academic work on structuring the dark web. Its founder has published several papers on the indexing methodology and on automated filtering of problematic content.
Catalogue and coverage
Ahmia indexes between 20,000 and 30,000 active .onion services at any given time, according to figures published by the team. The crawler follows links from known directories and previously indexed pages at a high update frequency. Inaccessible services are flagged and progressively removed from the index. The catalogue is smaller than Haystak's or Torch's, but the quality of entries is significantly higher.
Filtering policy
This is Ahmia's defining feature: active filtering of manifestly illegal content. The official blacklist excludes child sexual abuse material, terrorist propaganda, and a few other categories clearly condemned by international consensus. This list is publicly maintained on GitHub and can be expanded by the community through reports.
The filtering is not perfect — no automated filtering is — but Ahmia is by far the most rigorous .onion engine on this question. It is the only engine that the Tor Project explicitly recommends in its documentation.
Interface and features
Ahmia's interface is minimalist and efficient: a search field, results in a list, and a few basic filters (primary site language). No frills, no ads. Results display the page title, the full .onion address, a relevant text excerpt, and the approximate date of last access.
Ahmia is also accessible on the clearnet via ahmia.fi, making it a practical bridge between the two worlds. Researchers, journalists, and curious users can explore the .onion ecosystem without necessarily launching Tor Browser. Results always point to .onion addresses that need to be opened with Tor, but browsing the results themselves is possible without it.
Strengths and limitations
Strengths: clear ethical policy, active filtering, clean interface, accessible on both clearnet and .onion, recommended by the Tor Project, publicly identified and accessible founder, documented methodology.
Limitations: smaller catalogue (a few tens of thousands of entries), less useful for exhaustive searches, few advanced search operators, depends on a small team so vulnerable to shutdown if funding stops.
Torch: the comprehensive veteran
Torch (short for "Tor Search") is one of the oldest dark web search engines, operational since the early 2010s. Its catalogue is often cited as the most comprehensive for certain types of searches, at the cost of a complete absence of ethical filtering.
Catalogue and coverage
Torch claims a catalogue of several million indexed pages, making it one of the largest in the .onion ecosystem. In reality, this figure is misleading: it includes a large number of dead addresses, defunct services, and duplicate mirrors. The number of services actually accessible at any given moment is probably far more modest.
Filtering policy
Torch operates a no-filtering policy: all sites discovered by the crawler are indexed without distinction. This includes potentially illegal content, which is why we do not recommend Torch for exploratory browsing. For specific uses (academic research on the dark web, investigative journalism), this comprehensiveness can be an asset, but it requires heightened vigilance from the user.
Interface and monetization
Torch's interface is functional but dated. Most notably, Torch displays advertising banners at the top of search results — something rare on the dark web. The advertisers are mainly other .onion services (marketplaces, mixing services, hosting providers), and these ads often promote dubious services. Clicking on them is strongly discouraged.
Strengths and limitations
Strengths: large catalogue, long history, useful for certain searches that other engines miss, has operated for a long time without notable interruption.
Limitations: no ethical filtering whatsoever, many dead addresses in the index, ads for dubious services, variable result quality. Not recommended for beginners or for exploratory browsing.
Haystak: the largest catalogue
Haystak is another veteran dark web engine that claims an even larger catalogue than Torch: over one billion indexed pages by its own figures. As with Torch, this number should be taken with caution, but Haystak is undeniably one of the engines with the largest raw index.
Catalogue and coverage
Haystak's crawler aggressively follows every link it finds and archives content even when services become inaccessible. As a result, Haystak often contains content that no longer exists, which can be useful for historical research but frustrating for everyday use.
Filtering policy and model
Like Torch, Haystak does not practice systematic ethical filtering. The service also offers a paid model: a "premium" version accessible by subscription, providing advanced features (Boolean searches, sophisticated filters, CSV export). This business model makes Haystak interesting for cybersecurity researchers and journalists, but less relevant for casual users.
Strengths and limitations
Strengths: the largest raw catalogue, advanced features in the premium version, useful for historical research on the dark web.
Limitations: limited free version, many dead addresses, no ethical filtering, dated interface, useful features require a paid plan.
DuckDuckGo on .onion: anonymous clearnet search
DuckDuckGo deserves a place in this comparison, even though it is not strictly a dark web search engine. DuckDuckGo indexes the regular clearnet, but offers a .onion version that lets you conduct standard searches without revealing your IP to the search engine. It is the cleanest combination for searching for information on the clearnet from Tor.
History and positioning
DuckDuckGo was founded in 2008 by Gabriel Weinberg in Pennsylvania, with privacy as its main commercial argument. Its .onion version has been active since 2010, making it one of the oldest .onion services still operational. Tor Browser has used DuckDuckGo as its default search engine for several years, owing to its exemplary compatibility with Tor.
What it does and does not do
DuckDuckGo does not search .onion sites. Its index covers the clearnet; you need Ahmia or Torch to search specifically for dark web services. What DuckDuckGo provides is the ability to search the regular web from Tor, without Google's aggressive CAPTCHAs and without advertising tracking.
Behind the scenes, DuckDuckGo draws on several sources: Bing (Microsoft) for web results, Wikipedia, specialized APIs (weather, calculations, conversions), and its own crawler. Result quality does not match Google's but remains very good for most everyday searches.
Strengths and limitations
Strengths: no CAPTCHAs on Tor, no tracking, good result quality, perfect integration with Tor Browser, stable .onion version for over ten years.
Limitations: does not index .onion services (you need Ahmia or another specialized engine for that), results can be less relevant than Google on highly technical queries.
Other engines and alternatives
Beyond the four main engines, a few alternative engines are worth mentioning.
DarkSearch is a more recent engine targeting cybersecurity professionals and researchers specifically. It offers an API for automating searches and keyword alert features. The service is freemium.
Kilos (formerly Grams) is a search engine specialized in dark web marketplaces, with a price aggregator for the various products sold. Its use is obviously ethically problematic, and we do not recommend it.
Manual directories such as OnionDir, The Hidden Wiki, and a few others are not search engines per se, but they play an essential complementary role. Where engines index automatically, directories select and verify manually. Every entry in OnionDir is tested before publication and re-verified periodically.
Comparison table
A summary of the main engines on key criteria.
Effective search strategy
To get the most out of the ecosystem, here is a three-step strategy that combines several tools.
Step 1: define your need. Are you looking for a specific service (an encrypted email, an anonymous clearnet engine, a sharing tool) or exploring a topic (legitimate .onions, censored media, privacy tools)? The first question points toward a directory; the second toward a search engine.
Step 2: start with directories. Our OnionDir directory lists verified services organized by category: media, privacy tools, encrypted email, search engines, forums. For well-known major services (BBC, NYT, ProtonMail, SecureDrop), manual directories are more reliable than automated engines.
Step 3: complement with Ahmia. For more precise searches or to discover services not listed in directories, Ahmia is the best complement. Its ethical filtering and clean interface make it the everyday tool for most serious users. If Ahmia does not find what you need, Torch and Haystak can sometimes fill the gaps — with the required vigilance.
For clearnet searches. If you are on Tor and want to search the regular web, DuckDuckGo on .onion is the best option. Avoid Google, which regularly blocks Tor exit nodes and imposes CAPTCHAs.
Going further
This article is part of a complete series on the dark web. For installation and Tor access, see our dark web access guide. To understand the technical underpinnings of .onion addresses, our article on what a .onion link is covers the cryptography. To distinguish the dark web from the deep web, our article Dark Web vs Deep Web is the reference.
For an overview of the most relevant services, see our top 30 unusual and legitimate .onion sites. For curiosities and urban legends, our unusual FAQ and our pillar article on 50 debunked dark web myths. Our glossary rounds things out with precise technical definitions.