Onion hidden sites are websites with .onion addresses accessible only through Tor, providing end-to-end encryption from your browser to the host automatically (1). Unlike regular websites where exit nodes can see unencrypted traffic, onion services encrypt content throughout the entire connection: you connect through a rendezvous point that relays encrypted messages without being able to decrypt them (1), while the service itself remains hidden behind introduction points (2).
What Are Onion Hidden Sites and How Do They Differ from Regular Websites?
Onion hidden sites are websites with .onion domain addresses that you can access only through the Tor network. Unlike regular websites with .com or .org domains that resolve through standard DNS, .onion addresses are cryptographic identifiers that the Tor network uses to route traffic anonymously. You need Tor Browser or another Tor-enabled client to reach these sites—typing a .onion address into Chrome or Firefox will not work.
The defining technical difference is how encryption protects your connection. Onion service traffic is encrypted end-to-end from your client to the onion host automatically, providing strong SSL/HTTPS-like encryption built into the protocol (1). This means the content of your communication remains protected throughout the entire path, not just at the transport layer. When you visit a regular website through Tor, your traffic exits the Tor network at an exit relay before reaching the destination server. If that website does not use HTTPS, the exit relay can see the content of your requests and the data the server sends back. Onion services eliminate this vulnerability entirely—there is no exit relay in the connection, so no intermediate node can expose plaintext content.
Why This Matters for Privacy
Both you and the onion service build anonymous circuits through the Tor network and meet at a rendezvous point, which relays encrypted messages without being able to decrypt them (1). Neither the service nor any relay in the chain learns your IP address, and you cannot learn the service's IP address. This architecture makes onion sites a stronger choice for privacy-focused communication, censorship circumvention, and source protection. Organizations like ProPublica and the BBC operate onion versions of their sites to reach readers in censored regions or offer journalists an extra layer of anonymity.
The technology itself is neutral. It enables private, censorship-resistant communication without revealing IP addresses of either party (1). Onion sites are not inherently "dark web" locations—they are privacy-focused services that use cryptographic routing and end-to-end encryption to protect both the server and the visitor.
For practical guidance on accessing these services, see Onion Sites on Tor: What They Are and How to Access Them.
How Onion Service Connections Work: Circuits, Introduction Points, and Rendezvous
Connecting to an onion site involves a multi-step handshake that establishes anonymous bidirectional communication without revealing either party's IP address. The process uses three key components: the rendezvous point, introduction points, and anonymous circuits built by both client and service.
When you want to connect to a hidden service, your client first chooses a random Tor node to act as the rendezvous point and builds a circuit to it (2). At the same time, your client retrieves the hidden service descriptor—a document listing the service's introduction points—and builds a second anonymous circuit to one of those introduction points (2). The introduction point forwards your encrypted connection request to the service, which then builds its own anonymous circuit to the rendezvous point you selected.
The Rendezvous Point's Role
The rendezvous point serves as a relay between your circuit and the service's circuit. It passes end-to-end encrypted messages between the two parties without being able to decrypt the content (1). This node knows only that two anonymous circuits are exchanging data—it cannot identify you, cannot identify the service, and cannot read the messages. Because both circuits are already encrypted through multiple Tor relays, the rendezvous point sees only layers of encryption with no plaintext exposed.
This design ensures that no single relay in the path can link your IP address to the onion site you visit. The introduction points see only the service's circuit and an encrypted request; the rendezvous point sees only two circuits exchanging encrypted traffic. The multi-step handshake distributes trust across multiple nodes, none of which holds enough information to compromise your anonymity or the service's location.
For readers concerned about surveillance, this architecture removes the exit relay vulnerability present in regular Tor browsing—there is no point in the connection where an intermediate node can see unencrypted content.
Encryption Layers Protecting Onion Services
All connections in Tor use TLS link encryption so observers cannot look inside traffic (3). When you build a circuit through the Tor network, your client establishes ephemeral encryption keys with each relay in the path, creating multiple layers of encryption where each relay can decrypt only one layer to learn the next hop (3). This means your data is wrapped in three or more encryption layers as it travels through the circuit—each relay peels off one layer, sees only the next relay's address, and forwards the still-encrypted traffic onward.
Why Onion Services Eliminate the Exit Relay Risk
In regular Tor browsing, the final relay in your circuit—the exit node—decrypts the outermost layer and forwards your request to the destination website. If that website does not use HTTPS, the exit relay can read the plaintext content of your traffic (3). Onion services remove this vulnerability entirely. There is no exit relay in an onion service connection—both you and the service build anonymous circuits that meet at a rendezvous point, which relays end-to-end encrypted messages without being able to decrypt them (1). The ephemeral keys you establish with the service ensure that only your client and the onion host can read the content, not the rendezvous point or any relay in either circuit.
Two-Layer Encryption for Hidden Service Descriptors
Hidden service descriptors—documents listing the service's introduction points—are themselves protected by two layers of encryption (4). The first layer provides confidentiality against entities that do not know the public key of the hidden service, such as the HSDir nodes that store these descriptors (4). This prevents storage nodes from reading the introduction points or any metadata about the service. The second layer protects against entities that do not possess valid client credentials when restricted discovery is enabled (4). If a service operator configures client authorization, only users with the correct credentials can decrypt the second layer and retrieve the introduction points needed to connect.
This two-layer design means that even the Tor nodes responsible for storing and distributing hidden service descriptors cannot learn where the service's introduction points are located or who is authorized to connect. The encryption keys are derived from the service's cryptographic identity, ensuring that only the intended parties can access the information needed to establish a connection.
End-to-End Encryption: How Onion Services Protect Both Ends of Communication
Onion service traffic is encrypted end-to-end from your client to the onion host, providing strong SSL/HTTPS-like encryption automatically without requiring separate certificates (1). This means the content of your communication remains protected throughout the entire path—neither Tor relays nor the rendezvous point can see what you are sending or receiving. The encryption is built into the protocol, so you do not need to verify certificates or configure additional security layers when connecting to a .onion address.
What End-to-End Encryption Means in Practice
When you visit an onion site, your Tor Browser establishes ephemeral encryption keys directly with the onion host. These keys ensure that only your client and the service can decrypt the messages exchanged between you. The rendezvous point relays end-to-end encrypted messages between your circuit and the service's circuit without being able to decrypt the content (1). It sees only layers of encryption—no plaintext, no metadata about what you are accessing, and no information linking you to the service.
This differs fundamentally from regular Tor browsing. When you visit a normal website through Tor, your traffic exits the Tor network at an exit relay before reaching the destination server. If that website does not use HTTPS, the exit relay can read the plaintext content of your requests and the server's responses. Onion services eliminate this vulnerability entirely—there is no exit relay in the connection, so no intermediate node can expose unencrypted traffic.
Why This Architecture Matters for Privacy
The absence of an exit relay removes a common surveillance point. Exit nodes in regular Tor browsing can monitor unencrypted traffic, log visited domains, or inject malicious content into HTTP responses. Onion services close this gap by encrypting the entire communication path from your device to the service's server. Neither the rendezvous point nor any relay in your circuit or the service's circuit can decrypt the content (1).
This built-in end-to-end encryption also protects the service operator. The onion host does not learn your IP address, and you do not learn the service's IP address. Both parties remain anonymous to each other and to all relays in the network. For readers concerned about metadata collection or traffic analysis, this design ensures that no single node holds enough information to compromise your identity or the service's location.
To understand how this encryption applies when accessing specific onion sites, see Dark Web Sites on Tor Browser: Access and Safety Guide.
Onion Service Connection: Encryption Layers and Node Roles
| Connection Stage | Node Type | What Node Sees | Encryption Layer | Can Decrypt Content? |
|---|---|---|---|---|
| Client builds circuit | Guard relay | Client IP, next relay address | TLS + ephemeral keys | No |
| Client builds circuit | Middle relay | Previous relay, next relay address | TLS + ephemeral keys | No |
| Client chooses meeting point | Rendezvous point | Two anonymous circuits exchanging data | End-to-end encrypted messages | No |
| Service builds circuit | Introduction point | Service circuit, client request (encrypted) | TLS + ephemeral keys | No |
| Service builds circuit | Service-side relays | Previous relay, next relay address | TLS + ephemeral keys | No |
| Descriptor storage | HSDir node | Encrypted descriptor blob only | Two-layer descriptor encryption | No |
| Final connection | Onion host server | Decrypted content from client | End-to-end encryption terminates here | Yes |
| Regular Tor browsing (comparison) | Exit relay | Destination site, unencrypted HTTP content | TLS to exit, plaintext after | Yes (if no HTTPS) |
Q&A
- What are onion sites used for
Onion sites serve privacy-focused communication, censorship circumvention, and source protection. Organizations like ProPublica and the BBC operate onion versions of their sites to reach readers in censored regions or offer journalists an extra layer of anonymity. The technology itself is neutral—it enables private, censorship-resistant communication without revealing IP addresses of either party (1).
- How to find onion sites on Tor
You find onion sites through directories, search engines designed for .onion addresses, or direct links shared outside the Tor network. Your Tor client retrieves the hidden service descriptor from HSDir nodes, which store encrypted blobs listing the service's introduction points (4). Without a valid .onion address or directory entry, discovering services requires manual sharing or specialized onion search tools.
- Are onion sites safe to use
Onion sites provide strong encryption and anonymity by design—traffic is encrypted end-to-end from your client to the onion host, and no relay can decrypt the content (1). However, safety depends on the service operator's intent and your behavior: malicious sites can host scams or malware, and downloading files or disabling Tor Browser protections exposes you to risks. The protocol protects your connection, not the trustworthiness of the content.
- What is the difference between Tor Browser and a proxy server
Tor Browser routes your traffic through multiple relays with layered encryption, where each relay decrypts one layer to learn the next hop (3). A proxy server typically routes traffic through a single server that can see your original IP and the destination, offering no encryption unless combined with a VPN. Tor's multi-hop design and ephemeral keys prevent any single relay from linking you to your destination, while a proxy operator sees both ends of the connection.
- Can I be tracked while using Tor
Tor prevents relays and the rendezvous point from linking your IP address to the onion site you visit—traffic is encrypted end-to-end and the rendezvous point only relays encrypted messages without decrypting them (1). However, you can be tracked if you log into accounts tied to your real identity, disable JavaScript protections, or download files that execute outside Tor Browser. The protocol hides your network location, but behavioral mistakes or browser fingerprinting can still reveal your identity.
- Is Tor illegal
Using Tor and accessing onion sites is legal in most countries, including the United States. The technology itself is designed for privacy and censorship resistance, and organizations like news outlets operate onion services for legitimate purposes. However, using Tor to commit illegal activities—such as accessing prohibited content or conducting fraud—remains illegal regardless of the tool. The legality depends on your actions, not the software.
- What are some risks associated with using onion sites
Risks include encountering scams, malware, or illegal content hosted by malicious operators, since onion sites have no central authority verifying legitimacy. Even though connections are encrypted end-to-end (1), downloading files or enabling plugins can expose your real IP or compromise your device. Law enforcement monitors certain onion services, and behavioral mistakes—like reusing usernames or disabling Tor Browser protections—can link your activity to your identity despite the protocol's anonymity.
Key Takeaways
Onion sites use .onion addresses accessible only through Tor, with end-to-end encryption built into the protocol from your client to the onion host automatically (1).
No exit relay exists in onion service connections, eliminating the risk that an intermediate node can see plaintext content—the rendezvous point only relays encrypted messages without decrypting them (3) (1).
Hidden service descriptors are protected by two layers of encryption: the first prevents HSDir storage nodes from reading introduction points, the second restricts access to clients with valid credentials when enabled (4).
Your IP address and the service's IP address remain hidden from each other and from all relays, as both parties build anonymous circuits that meet at a rendezvous point which cannot identify participants (1).
Safety depends on operator intent and your behavior—the protocol protects your connection, but malicious sites can still host scams or malware, and disabling Tor Browser protections exposes you to tracking.
To explore how these encryption mechanisms apply when accessing onion sites in practice, see Dark Web Sites on Tor Browser: Access and Safety Guide.
References
[1] Tor Project | How do Onion Services work?
[2] Protocol overview - Tor Specifications
[3] Tor cryptographic keys - How Tor works - About Tor — Tor
[4] Hidden service descriptors: encryption format [HS-DESC-ENC] - Tor Specifications
Where to look next
These services are useful starting points for further research. Tools & services.