Torch Search Engine: Navigating the Dark Web Safely

What Torch Actually Is (and What It Is Not)

Torch is a search engine that runs as a Tor hidden service, meaning you access it through a .onion address inside Tor Browser or another Tor-routing application[1]. It crawls publicly reachable .onion sites—the hidden services that make up part of the dark web—and returns links when you search[2][7]. Torch has operated since roughly 2012, making it one of the longest-running tools of this kind[3].

What Torch does not do

Torch is not a web browser. A separate, unrelated project called Torch Browser is a Chromium-based clearnet browser with no connection to the Tor network; searching "Torch browser" will often return that product instead of the onion search engine.

Torch does not index the entire dark web. An .onion address is a cryptographic identifier derived from a service's public key, not a registered domain name, so no central list exists[7]. Torch starts from a seed list of known addresses, fetches each page, and follows every .onion link it finds; directories such as The Hidden Wiki and forum posts form the backbone of its coverage[7]. If a site is never linked publicly, Torch will not discover it.

Torch does not host, verify, or moderate the sites it indexes[5]. It returns whatever matches your query, including defunct links, scams, and illegal material. The engine performs little or no content filtering, unlike Ahmia, which blocks known child-abuse domains and publishes a ban list[4].

What you need to use it

You must route traffic through Tor. Visiting the clearweb mirror torch.cx in a standard browser will show a landing page but will not let you follow .onion results[1]. Install Tor Browser, navigate to the official onion address, and search from there[1]. Researchers have found that most links claiming to be Torch are malicious clones designed to redirect you to phishing or scam sites[4], so bookmark the verified address and check it matches before you type a query.

Official Access and How to Verify the Link

Torch has no clearnet version that works in a standard browser. You must reach it through its .onion address inside Tor Browser[1]. The official address documented on Torch's own site is torchsfe235y6d7wguqo6g4ucxqkq7frrm5fpgkjssdhthsq4kjmmisid.onion[1]. The clearweb mirror torch.cx will load a landing page in any browser, but clicking search results will fail because .onion links require Tor routing[1].

Why verification matters

Researchers testing dark-web directories found that most links labelled 'Torch' were malicious clones returning fewer than 150 addresses and redirecting visitors to phishing pages[4]. Because an .onion address is a cryptographic hash rather than a registered domain, no central authority publishes a definitive list[7]. You cannot look up 'Torch' in a registrar database the way you would torch.com.

Before you visit any .onion address claiming to be Torch, cross-check it against at least two independent sources:

• The official Torch site itself (torchsfe235y6d7wguqo6g4ucxqkq7frrm5fpgkjssdhthsq4kjmmisid.onion)[1].

• Community-maintained directories such as The Hidden Wiki or curated lists on Reddit.

• Tor Project community forums and wikis, where volunteers share known-good addresses.

Write down or bookmark the verified address after your first successful visit. If the address you find elsewhere differs by even one character, treat it as a phishing attempt and do not proceed.

What a lookalike domain looks like

A fake onion address may swap a single letter—'torchsfe235y6d7wguqo6g4ucxqkq7frrm5fpgkjssdhthsq4kjmmisid' becomes 'torchsfe235y6d7wguqo6g4ucxqkq7frrm5fpgkjssdhthsq4kjmmizid'—or use a completely different hash that simply includes the word 'torch'. Always compare the full string, not just the recognisable fragment.

How Torch Indexes Onion Sites

Torch builds its database by crawling publicly accessible .onion pages, much like a clearnet search engine follows links across the web[7]. It starts from a seed list of known addresses—directories, forums, and link collections such as The Hidden Wiki—fetches each page, and extracts every .onion link it finds[7]. Those new addresses are added to the queue, and the process repeats. If a hidden service is never linked publicly, Torch will not discover it.

What the crawler cannot reach

Many hidden services remain outside Torch's index because they block crawlers, require authentication, or operate by invitation only. A private forum that checks for a session cookie before serving any content will return an empty page to Torch's bot. A marketplace that hides its product listings behind a login screen will appear as a single landing page with no deeper links. An .onion site that publishes its address only in encrypted messages or private chat rooms will never enter the seed list in the first place.

Because an .onion address is a cryptographic identifier derived from the service's public key rather than a registered domain, no central registry exists to enumerate every hidden service[7]. Torch can only index what it can find and fetch.

The index is not real-time

Torch's crawler visits each known address periodically, but the interval between visits can range from days to weeks. A newly created onion forum may not appear in search results until the crawler discovers a link to it on another indexed page and then schedules a fetch. If the forum goes offline before that fetch completes, it will never be indexed. If a site changes its content or moves to a new address, Torch's results will reflect the old snapshot until the next crawl.

An Austrian study found that Torch returned 4,190 addresses during the observation period—the lowest count among six sources tested—while Ahmia's address list ranged between 18,000 and 22,000 entries across a single week[4]. Torch's often-cited claim of having indexed a billion pages comes from Torch itself and has never been independently verified[4]. Treat the index as a starting point, not a complete map of the dark web.

What You Can and Cannot Find on Torch

Torch indexes publicly linked .onion pages, so you will see directories, forums, marketplaces, whistleblowing platforms, privacy-tool mirrors, and informational sites that other hidden services reference[7]. If a page appears in The Hidden Wiki, a forum thread, or a link list and Torch's crawler can fetch it, the page enters the index. That coverage means you can search for onion versions of well-known services, topic-specific discussion boards, and archived documents that clearnet search engines do not reach.

What Torch does not index

Torch cannot crawl password-protected sites, private databases, or invitation-only forums[7]. A marketplace that hides product listings behind a login screen will appear as a single landing page with no deeper links. A whistleblowing platform that requires a passphrase before serving any content will return an empty page to the crawler. Many illegal marketplaces actively block automated bots to avoid law-enforcement mapping, so they will not appear in Torch results even if the onion address is known.

Because an .onion address is a cryptographic identifier rather than a registered domain, no central list exists to enumerate every hidden service[7]. If a site is never linked publicly—shared only in encrypted messages or private chat rooms—Torch will not discover it.

Treat every result as untrusted

Torch performs little or no content filtering, unlike Ahmia, which blocks known child-abuse material[4]. A search for any term may return defunct links, scam sites, or honeypots operated by law enforcement. Researchers found that most links claiming to be Torch were malicious clones designed to redirect visitors to phishing pages[4], and the same risk applies to any onion address in Torch's results.

Before you visit a site from Torch, cross-check the address against community-maintained directories or forums. Bookmark verified addresses after your first successful visit. Disable JavaScript in Tor Browser's security settings, and never enter personal information or credentials on an onion site you have not independently verified. Some results will load an error page or time out; others will serve malware or attempt to de-anonymise you. Torch does not verify, moderate, or endorse any indexed content[5], so the responsibility to assess each link rests entirely with you.

Safety and Content Considerations

Torch does not filter, curate, or moderate the sites it indexes[5]. A search for any term may return links to disturbing material, scams, defunct pages, or content that is illegal to access in your jurisdiction. The engine performs little or no content filtering, unlike Ahmia, which blocks known child-abuse domains and publishes a ban list[4]. Torch states explicitly that it does not host, store, verify, or endorse third-party content[5], so the responsibility to assess each result rests entirely with you.

Using Torch is legal; accessing certain content is not

In most jurisdictions—including the United Kingdom—running Tor Browser and searching an onion index is lawful. The Tor network itself is a privacy tool used by journalists, activists, and researchers, and Torch is simply a search interface. However, clicking through to sites that host child sexual abuse material, sell controlled substances, or facilitate fraud is illegal regardless of how you found the link. Law enforcement agencies monitor known illegal marketplaces and may operate honeypots—fake onion sites designed to log visitor activity. Visiting an indexed page does not grant immunity from prosecution if the content violates UK law.

Check directory reputation before clicking unknown links

Before you visit any result from Torch, cross-check the onion address against community-maintained directories such as The Hidden Wiki or curated lists on Reddit. If the address appears in multiple independent sources and has user comments describing the service, the risk of encountering a scam or honeypot drops. If the address is absent from every directory you check, treat the link as untrusted and do not proceed. Researchers found that most links claiming to be Torch were malicious clones designed to redirect visitors to phishing pages[4]; the same risk applies to any onion address in Torch's results.

Disable JavaScript and avoid downloads

Open Tor Browser's security settings and select 'Safest', which disables JavaScript, limits fonts, and blocks some media. Many de-anonymisation attacks rely on JavaScript to fingerprint your browser or leak your real IP address through WebRTC. Even a site that appears benign may serve malicious scripts. Never download files from onion sites unless you have verified the source through multiple independent channels; executable files and PDFs can contain malware designed to bypass Tor routing or exploit browser vulnerabilities.

Never enter personal credentials on onion sites

Do not log in to clearnet accounts—email, social media, or payment services—while Tor Browser is running. Do not create accounts on onion sites using your real name, existing email address, or any identifier tied to your offline identity. If a site asks for payment, assume the transaction is irreversible and that the operator may disappear without delivering goods or services. Torch's stated logs contain no IP addresses[6], but the sites you visit after leaving Torch may log everything you submit.

Typical Mistakes When Using Torch

First-time users often make avoidable errors that expose them to scams, malware, or de-anonymisation. Below are the most common mistakes and a one-sentence correction for each.

Accessing Torch through a clearnet proxy

Some websites offer to 'browse .onion sites without Tor' by acting as a proxy between your browser and the Tor network. When you use such a service, the proxy operator sees every search query you submit, every result you click, and your real IP address—defeating the anonymity that Tor provides. Always access Torch directly through Tor Browser at its official onion address[1], never through a third-party clearnet gateway.

Assuming Torch equals full dark web coverage

Torch indexes only publicly linked .onion pages that its crawler can fetch[7], so private forums, password-protected marketplaces, and invitation-only services will not appear in results. An Austrian study found that Torch returned 4,190 addresses during the observation period—the lowest count among six sources tested[4]. Treat Torch as one starting point among several; cross-reference results with other dark web search engines and community directories to broaden your coverage.

Clicking results without verifying the onion address

Researchers found that most links claiming to be Torch were malicious clones designed to redirect visitors to phishing pages[4], and the same risk applies to any address in Torch's results. Before you visit a site, cross-check the full .onion string against at least two independent sources—community-maintained directories, forum threads, or Tor Project wikis—and bookmark verified addresses after your first successful visit.

Enabling scripts or downloading files from unknown onion services

Many de-anonymisation attacks rely on JavaScript to fingerprint your browser or leak your real IP address through WebRTC. Open Tor Browser's security settings and select 'Safest' to disable JavaScript, limit fonts, and block some media; never download executable files or PDFs from onion sites unless you have verified the source through multiple independent channels.

Confusing Torch with the Torch browser

'Torch' is the name of a search engine[1]; it is not a web browser and does not replace Tor Browser. You must download Tor Browser from the Tor Project's official site and use it to visit Torch's onion address[1]. Any software claiming to be 'Torch Browser' is either a scam or a rebranded fork with unknown modifications; do not install it.

How to Verify That a Torch Result Is Still Active

Many onion services disappear within days or weeks of launch, so a dead link in Torch's results is common and does not indicate that your connection has been compromised. Hidden services go offline when operators shut them down, when hosting infrastructure fails, or when law enforcement seizes servers. Because Torch's crawler may take days to revisit a page[7], the index often contains addresses that no longer resolve. Below is a simple workflow to check whether a result is still reachable before you spend time exploring it.

Copy the full onion address and cross-check it

Select the entire .onion string from Torch's result page—for example, examplexyz123456789abcdefghijklmnopqrstuvwxyz.onion—and paste it into a text editor. Open a second directory such as Ahmia or a curated list on Reddit, then search for the same address. If the address appears in multiple independent sources with recent user comments describing the service, the site is more likely to be active. If no other directory lists the address, treat it as high-risk and proceed with caution.

Open the address in a fresh Tor Browser session

Launch Tor Browser, paste the onion address into the location bar, and press Enter. Wait up to sixty seconds; onion services can take longer to resolve than clearnet sites because Tor must build a circuit and negotiate an introduction. If the page loads, check for a valid onion service descriptor by clicking the padlock icon in the address bar. A green padlock with 'Onion Site' confirms that Tor Browser has established an authenticated connection to the hidden service. If you see a certificate warning or no padlock, the site may be a phishing clone.

Expect dead links and move on

If the connection times out or returns 'Unable to connect', the service is offline. Close the tab, return to Torch, and try the next result. Do not refresh repeatedly or attempt to troubleshoot the connection; a non-responsive onion address usually means the operator has shut down the service, not that your Tor circuit is broken. Bookmark any address that loads successfully so you can return directly without searching again.

Torch Compared With Other Onion Search Tools

Torch is one of several search engines designed for .onion addresses, and each tool makes different trade-offs between index size, content filtering, and accessibility. Understanding these differences helps you choose a starting point that matches your needs and risk tolerance.

Torch: large index, minimal filtering

Torch claims to maintain the largest index of onion services[2], though its often-cited figure of one billion pages has never been independently verified[4]. The engine performs little or no content filtering[4], so search results may include defunct links, scam sites, or illegal material. An Austrian study found that Torch returned 4,190 addresses during the observation period—the lowest count among six sources tested[4]—suggesting that its crawler may prioritise breadth over freshness. Torch offers a clearnet mirror at torch.cx[1], but accessing it through a proxy defeats the anonymity that Tor provides; always use the onion address inside Tor Browser.

Ahmia: filtered index with clearnet access

Ahmia was created by Finnish security researcher Juha Nurmi during Google Summer of Code in 2014 with support from the Tor Project, and it remains open source[4]. Unlike Torch, Ahmia blocks known child-abuse material and publishes its list of banned onion services as hashes[4]. Ahmia's address list ranged between 18,000 and 22,000 entries across a single week in the same Austrian study[4]—roughly four to five times the count observed from Torch. Ahmia also provides a clearnet mirror that lets you search onion addresses without running Tor Browser, though the mirror operator can see your queries and IP address. If you want a filtered index and are willing to accept a smaller result set, Ahmia is a safer starting point than Torch.

DuckDuckGo Onion: privacy-focused clearnet search via Tor

DuckDuckGo operates an onion mirror of its clearnet search engine, accessible through Tor Browser. The service does not index .onion addresses; instead, it searches the same clearnet web that you would reach from any browser, but routes your queries through the Tor network to protect your IP address. If you need to search for information on the public web without leaving Tor Browser, DuckDuckGo Onion is the correct tool. If you need to find hidden services, it will not help.

Haystak: onion-only indexing with ads

Haystak focuses exclusively on indexing .onion sites and displays advertisements to fund its operation. The engine does not publish detailed statistics about its index size or crawler behaviour, so independent comparisons are difficult. Haystak's results overlap with Torch and Ahmia, but the presence of ads increases the risk of malicious scripts or phishing links embedded in the page. If you choose Haystak, set Tor Browser's security level to 'Safest' to block JavaScript before you load any search results.

Which tool to start with

If you need the widest possible set of results and are prepared to verify every link manually, use Torch. If you want a smaller, filtered index that blocks known abuse material, use Ahmia. If you need to search the clearnet without leaving Tor Browser, use DuckDuckGo Onion. Cross-reference results from multiple engines and check addresses against community-maintained directories before you visit any onion site. No single search engine covers the entire dark web, and no index can guarantee that a result is safe or legal to access.

Common Mistakes and Misconceptions

Most people approach Torch with assumptions carried over from clearnet search engines like Google or Bing. Those assumptions break down quickly on the Tor network, where indexing is partial, addresses are cryptographic rather than human-readable, and verification is entirely manual. The four misconceptions below cover the errors we see most often among new users.

Treating Torch as a complete index of the dark web

People assume that a search engine lists everything that exists, so they type a query into Torch and expect every relevant hidden service to appear in the results. Torch grows its index by crawling from a seed list and following .onion links it finds, which means it only sees pages that are publicly linked from directories, forums, or other indexed sites[7]. Private forums, password-protected marketplaces, and invitation-only services never appear because no public link points to them. An Austrian study observed 4,190 addresses from Torch during the research period—the lowest count among the six sources tested[4]—so the gap between what exists and what Torch shows is substantial. Use Torch as a starting point, then cross-reference with other directories and engines rather than treating it as a definitive map.

Believing Torch's self-reported index size

Torch describes itself as the largest index of onion services and claims to have indexed over a billion pages, but that figure comes from Torch itself and has never been independently verified[4]. The same study that observed 4,190 addresses from Torch found Ahmia's list ranged between 18,000 and 22,000 entries in a single week[4], which is hard to reconcile with a billion-page claim. The discrepancy matters because a larger claimed index can create false confidence that a search has been exhaustive. Judge Torch by what it actually returns for your queries, not by the marketing number on its About page.

Confusing Torch the search engine with Torch the browser

Several software projects have used the name 'Torch', including a Chromium-based browser that has nothing to do with the Tor network. Torch in this context is a search engine that indexes .onion addresses and requires Tor Browser to access them[1]. If you search for 'Torch download' on the clearnet, you may end up installing an unrelated browser, or worse, a malicious clone. Researchers found that most links claiming to be Torch were fake clones designed to redirect users to scam sites[4]. Get Tor Browser only from the Tor Project's official site, then visit Torch's onion address listed on its own site[1].

Assuming every result in Torch is safe because it is indexed

Indexing is not endorsement, and Torch's Terms of Service state plainly that the engine does not host, store, mirror, or control third-party content and does not endorse, verify, or moderate external content[5]. A result appearing in Torch means only that the crawler found the page at some point; it says nothing about whether the site is legitimate, safe, or still active. Because Torch performs little or no content filtering[4], results may include scam sites, phishing clones, or illegal material. Verify every address against at least two independent sources before you visit, and keep Tor Browser's security level set to 'Safest' to block scripts that could fingerprint or de-anonymise you.

Expecting Torch to work reliably through a clearnet mirror

Torch publishes a clearnet mirror at torch.cx[1], and some users try to access it from a regular browser to avoid installing Tor Browser. When you do that, the mirror operator sees your real IP address, every query you submit, and every result you click—the exact opposite of what Tor is designed to protect. The mirror exists for convenience, not for anonymity, and using it defeats the purpose of searching the dark web safely. Always open Torch's onion address inside Tor Browser[1]; treat the clearnet mirror as a fallback for checking whether the service is up, never as your primary access method.

Conclusions

• Treat Torch as a partial index, not a vetting service: a listing means a crawler found a page, not that the address is safe or active[5][4].

• Open Torch only through Tor Browser and its .onion address[1]. The clearnet mirror at torch.cx reveals your real IP and queries to the operator[1].

• Before you click any result, check the .onion address against a second directory. Torch does not reliably screen out scam, phishing, or illegal pages[4].

• Set Tor Browser’s security level to ‘Safest’ before any result loads, so scripts cannot run.

Next, compare Torch with Ahmia and filtered directories using the Best Dark Web Search Engines for 2026 checklist before your first search.