Oct 3, 2026 · Protocols · About 10 min read

Shadowrocket Supported Protocols: Shadowsocks, VMess, VLESS, Trojan, WireGuard and Hysteria2

See which protocols Shadowrocket supports, from Shadowsocks and VMess to WireGuard and Hysteria2, and what to enter under Add Server.

In this guide

Shadowrocket supports a range of proxy and tunneling protocols. When adding a connection, first match the protocol to the service details you already have, then check the address, port, authentication and transport settings. This guide is for users with connection details who need help filling in Add Server. It explains how to distinguish protocols, TLS and routing rules, then import a configuration and troubleshoot in a consistent order. A protocol name alone cannot predict speed or replace the server-side settings.

First, distinguish protocols, transports and routing rules

A protocol defines how the client communicates with a server. For example, Shadowsocks uses a specified encryption method and password, while WireGuard uses keys to establish a tunnel. Transport settings define how the connection is carried; some VMess and VLESS configurations, for instance, also use WebSocket and TLS. Routing rules determine which outbound route a destination uses. Check these three layers separately: enabling TLS does not change the protocol, and editing a rule will not fix an incorrect password.

3 layers
Check the protocol, transport settings and routing rules separately
443
HTTPS commonly uses this port, but entering it does not guarantee a connection
2 ways to get the app
iPhone / iPad: $2.99 one-time purchase; Mac / Windows / Android / Linux: free

The steps below follow Home → “+” in the top-right corner → Add Server. The fields shown vary by protocol, and some advanced options depend on the platform and the editor currently open. Use the actual Type list and editing screen as your reference. Do not select a different protocol just to make the fields look the same. This guide covers common types, not every available option.

Field guide to Shadowsocks, VMess, VLESS and Trojan

When adding a connection manually, Address is the connection destination and Port is the port the server actually listens on. Authentication and transport details in the table must come from your existing configuration; an address and port alone are usually not enough to establish a working connection. Do not guess the Type from the server’s location, name or port number.

Details to prepare and check for common protocols
Protocol Type Information to prepare Check first
Shadowsocks Address, Port, Password and the corresponding encryption method The encryption method must match. Any plugin or additional parameters must also match the server configuration.
VMess Address, Port, UUID, and the transport and security settings specified in the configuration UUID, transport type and TLS status. A significantly incorrect device clock may also affect authentication.
VLESS Address, Port, UUID, and the transport and security settings specified by the server Do not copy VMess authentication options. Enter TLS and other security settings according to the original configuration.
Trojan Address, Port, Password, and TLS parameters such as the server name Check that the password and certificate name match. Do not assume the server IP is the name listed in the certificate.

Shadowsocks: both the password and encryption method must be correct

For Shadowsocks, Password and the encryption method together determine authentication and how data is handled. Even if Address and Port establish a network connection, an incorrect encryption method can still prevent data from being transferred. Do not replace a method just because its name seems newer; changing it on the client alone will not make it compatible if the server configuration has not also changed.

VMess and VLESS: a UUID does not replace transport settings

VMess and VLESS can both use a UUID, but they are different Types. VLESS does not provide content encryption by itself; actual security depends on the accompanying security layer and deployment. If the original configuration uses WebSocket, check the path and Host as well as the UUID. If TLS is enabled, check the server name too. Paths are usually case-sensitive, and omitting the leading slash may send the request to the wrong location.

Trojan: resolve certificate errors before checking higher-level requests

Trojan is commonly used with TLS. The connection address, TLS server name and domain covered by the certificate can serve different purposes and are not interchangeable. If the logs report a certificate name or validity issue, check the device clock, server name and certificate status. Disabling certificate verification should not be a routine fix.

HTTP, SOCKS5, WireGuard and Hysteria2 compared

These connection types cannot be ranked by speed based only on whether they are encrypted. HTTP and SOCKS5 are common proxy protocols; WireGuard is an encrypted tunnel protocol carried over UDP; Hysteria2 is based on QUIC and also requires UDP access to the server. Networks handle TCP and UDP differently, so results may change when the same device switches between Wi-Fi and cellular.

Transport characteristics of other common types
Type Connection characteristics Information to check
HTTP / HTTPS An HTTP proxy does not mean the connection between the client and proxy server already uses TLS. An HTTPS proxy can use TLS for that connection. Proxy address, port, optional username and password, and whether the server requires TLS.
SOCKS5 The protocol does not provide content encryption. UDP availability also depends on the server and related settings. Address, port, authentication method, and whether UDP forwarding is available.
WireGuard Uses key-based authentication and carries the encrypted tunnel over UDP. Local private key, peer public key, Endpoint and tunnel address. Check the preshared key too if the configuration includes one.
Hysteria2 Based on QUIC. If UDP is blocked, changing routing rules alone cannot restore the connection. Address, port, authentication details, TLS server name and any other parameters specified in the original configuration.

In WireGuard, Endpoint is the peer’s public connection address and port; the tunnel address is used inside the tunnel. They are different fields. Private Key is the local identity credential, while Public Key identifies the peer; do not interchange them. If only some destinations are unreachable after import, check the route scope, DNS and MTU before generating new keys.

Bottom line: check network conditions before adjusting advanced settings

If WireGuard or Hysteria2 works on one network but times out after switching networks, first check whether the new network allows UDP traffic to the server port, then check the server status. Change one parameter at a time and record the result. Otherwise, changing the port, TLS and routing together makes it harder to identify the cause.

Add a server with Add Server or import an existing subscription

If you already have a complete subscription, preserve its protocol and transport settings to avoid manual entry errors. A subscription URL may contain working access credentials, so protect it like a password. This site does not provide, sell or recommend nodes or subscriptions; the steps below are only for connection details you already have.

  1. Prepare the settings

    Gather the original Type, Address, Port and authentication details. If TLS or WebSocket is used, note the server name, Host and path as well; do not copy only the server’s display name.

  2. Open the add screen

    Go to Home → “+” in the top-right corner → Add Server → Type. For manual setup, select the protocol specified in your connection details; do not guess based on the port or name.

  3. Enter or import the details

    For manual setup, fill in each field on the editor for the selected protocol. To use a subscription, choose Subscribe under Type, enter your existing subscription URL in URL, and save. The example address https://example.com/sub?token=xxxx illustrates the format only and cannot be used to obtain a connection.

  4. Check the result

    Return to Home and check the imported entry’s protocol type and group. Check the fields again after updating a subscription. Local manual edits may be overwritten by the original subscription configuration, so the number of entries alone does not confirm that an update worked correctly.

  5. Connect

    On Home, select the entry to test and check the current Global Routing mode before connecting. If the system asks to add a VPN configuration, review the prompt and allow it. Then visit a known working destination to verify that traffic is actually passing through.

For the first test, do not add automatic connection triggers. Settings → On Demand configures connections triggered by conditions; it is not a protocol authentication switch and cannot fix UUID, password or certificate errors. After verifying a manual connection, configure On Demand if needed, then test behavior when switching networks, disconnecting and reconnecting.

Check Global Routing after confirming the protocol

A server entry determines how to connect to the proxy endpoint; Global Routing determines whether requests use that endpoint. Diagnose these as separate issues. Config applies the current configuration’s rules; Proxy sends eligible traffic through the proxy; Direct connects directly. If you are using a Scene, check which settings it actually references rather than inferring the route from its name.

Config: check the rules before use

Recommended

The active configuration’s rules and policies determine the route. Use this mode for everyday traffic only if the configuration matches your expectations. The handling of unmatched requests depends on the actual configuration, including rules such as FINAL; it is not always direct.

Best for: everyday use when you have checked the configuration and want different destinations handled according to its rules.

Proxy: isolate the effect of routing rules

Temporarily use proxy mode to test whether the selected connection can carry requests. This can help narrow down rule-related issues, but it does not guarantee that all system traffic will use the proxy or bypass server authentication errors.

Best for: troubleshooting when an entry is selected but you need to check whether current rules are affecting it.

Direct: compare with a direct connection

Access the destination directly. If the webpage opens, that only confirms the destination is reachable over the current direct connection; it does not prove that a Shadowsocks, Trojan or other proxy entry is configured correctly.

Best for: checking basic connectivity and comparing direct access to the same destination.

Rule keywords on the Config page and protocol Types belong to different layers. The syntax below illustrates matching logic; it is not a complete file to replace an existing configuration. PROXY must refer to a usable policy in the active configuration.

DOMAIN-SUFFIX,example.com,PROXY
IP-CIDR,192.168.0.0/16,DIRECT,no-resolve
GEOIP,CN,DIRECT
FINAL,DIRECT

FAQ: importing a configuration does not guarantee traffic will pass

A successful import only means the client can read the configuration format. A latency test returning a result does not confirm that authentication, DNS, TLS and destination access are all working. Troubleshooting in the order “fields → network → mode → actual request” preserves useful clues better than repeatedly switching protocols.

Why does the webpage still not open when a latency result appears?

First check how the test works, then go to Home and verify the selected entry and Global Routing. A latency test may not cover a full webpage request. Temporarily use Proxy to test the same destination again, and check for authentication, TLS and DNS errors. Restore the previous mode when finished.

Can I just change VLESS to VMess and try again?

Changing the Type alone will not work. On Home, open the entry’s editor and check the UUID, transport type and security settings against the original configuration. The server must support the corresponding protocol; renaming it in the client does not make the server support a different authentication flow.

Hysteria2 keeps timing out. Should I change the port to 443?

Keep the original port for now. Test on another working network and check which port the server is listening on. Hysteria2 requires UDP access; entering 443 does not automatically provide HTTPS-style TCP connectivity. Change the client only if the server actually offers a new port.

Why did my manual changes revert after updating a subscription?

Open the updated entry and compare its fields with the original connection details you saved. Entries managed by a subscription may be regenerated during updates, so first check that the subscription source has the correct settings. To test independently, create a separate manual entry and give it a clear name to distinguish it from the subscription entry.

Will switching protocols always make the connection faster?

No. Keep the destination, network and test time consistent, then check whether the connection works, whether the first request completes and whether sustained transfers remain stable. Protocol is only one factor; server load, routing and packet loss matter too. Do not treat a single latency result as a long-term conclusion.

Download Shadowrocket