Shadowrocket supports multiple proxy protocols and WireGuard tunnel configurations, but changing Type alone cannot turn an existing connection into a different protocol. This guide is for users who already have their own provider and subscription. It walks through identifying the protocol, checking the fields, importing the configuration and testing the connection. First make sure the client settings match the server, then check Global Routing and your rules so configuration errors aren’t mistaken for protocol performance issues.
1. What protocols, transports and rules control
Shadowsocks, VMess, VLESS, Trojan, HTTP, SOCKS5, WireGuard and Hysteria2 are all covered here, but they don’t operate at exactly the same technical layer. WireGuard is a UDP-based tunnel protocol; HTTP and SOCKS5 define proxy communication methods; the other protocols each have their own authentication and data encapsulation. When adding a connection, use the type specified by your provider—not one that merely has a similar name.
TCP, UDP, WebSocket, TLS and QUIC are not interchangeable connection types. For example, a VLESS configuration may include both transport and TLS settings; Hysteria2 is based on QUIC and requires UDP connectivity. Even if Address and Port are identical, a mismatch in authentication, transport or TLS settings can prevent a connection from being established.
Key takeaway: match the server settings first
Test the connection once with the original imported settings, then adjust local options one at a time. Don’t change Type, Port, TLS and routing mode all at once; if the connection fails, you won’t know which change caused it.
2. Field checklist for eight protocols
Use this table to review an existing configuration, not as a set of server settings to copy. The fields shown in the editor vary by platform, selected Type and configuration. If a field isn’t shown, don’t put its value in Password or a notes field. If an imported subscription is missing a server-required parameter, first check the subscription content and the corresponding feature in your current client.
| Type | Fields to check | Common mistakes |
|---|---|---|
| Shadowsocks | Address, Port, encryption method, Password and any additional parameters explicitly required by the server. | Matching the password but not the encryption method; mistaking a service name for a protocol type. |
| VMess | Address, Port, UUID, encryption options, and parameters such as transport, TLS and path. | Copying the UUID but overlooking the WebSocket path or TLS hostname. |
| VLESS | Address, Port, UUID, and server-specified transport, security layer and Flow settings. | Assuming that a matching UUID means other VMess settings will work too. |
| Trojan | Address, Port, Password, TLS hostname and certificate validation settings. | When connecting by IP, replacing the certificate’s domain name with the IP address. |
| HTTP | Proxy server address, port, authentication details and whether to use TLS for the proxy connection. | Assuming that visiting an HTTPS website means the client-to-proxy connection is encrypted. |
| SOCKS5 | Server address, port, authentication method and whether the server supports UDP forwarding. | Assuming SOCKS5 encrypts all proxy traffic by itself. |
| WireGuard | Local Private Key, tunnel address, peer Public Key, Endpoint, Allowed IPs and more. | Confusing the local private key with the peer’s public key, or omitting the tunnel address and routing scope. |
| Hysteria2 | Address, Port, authentication details, TLS hostname and any obfuscation parameters required by the server. | Checking only the TCP port without confirming UDP connectivity; using settings for a different protocol. |
3. Shadowsocks, VMess and VLESS: authentication is only the first step
The Shadowsocks encryption method must match the server. Different methods may require different key material; don’t replace the password in the original configuration with a shorter, easier-to-remember string, or substitute an encryption method just because its name looks similar. If your provider supplies a plugin or obfuscation settings, confirm that the client supports that combination. The server address and password alone may not be enough.
VMess and VLESS commonly use a UUID as an identity field, but that doesn’t make them compatible. VMess also has its own authentication and encryption mechanisms; the VLESS protocol itself doesn’t encrypt data, so understand connection protection in the context of the security layer used by the configuration. TLS and transport settings must match the server.
- Check Address and TLS hostname separately: The connection target can be a domain name or an IP address. The hostname used for TLS name validation must match the server certificate and configuration. They don’t have to be the same.
- Preserve WebSocket parameters exactly: For example, if the server path is
/edge, don’t change it to/. Host, path and Address are separate fields. - Copy the complete UUID: Don’t include leading or trailing spaces, line breaks or non-ASCII punctuation. If your provider issues new credentials, update the actual connection configuration—not just its notes.
- Keep the device clock accurate: VMess authentication requires time synchronization, and TLS certificate validity checks also depend on the device clock. Correct any significant time discrepancy first.
4. Trojan, HTTP and SOCKS5: authentication is not encryption
Trojan typically uses TLS to establish a connection. Password handles authentication, while TLS provides the corresponding transport protection. A correct password can’t make up for an incorrect certificate name. If you see a certificate error, check the device clock, Address, TLS hostname and server certificate status rather than routinely disabling certificate validation.
Using an HTTP proxy and visiting a website over HTTPS are two separate things. A client can use an HTTP CONNECT request to create a tunnel to the destination; that doesn’t automatically mean the client-to-proxy connection uses TLS. SOCKS5 likewise provides no encryption by itself, and username-and-password authentication doesn’t mean the traffic is encrypted.
- Port: 443, 1080 and 8080 don’t identify the protocol
- Port 443 is often used for TLS services, 1080 is common for SOCKS, and 8080 is common for HTTP proxies—but servers can use other ports. Use the port in your existing configuration. Don’t infer Type just because a port looks familiar.
- If an HTTP proxy requires authentication, check the username and password separately. Don’t combine them in the server address field.
- A successful TCP connection to a SOCKS5 server doesn’t mean UDP forwarding works. Applications that rely on UDP also need support along the entire path, including the server and client.
- Understand the limits of an unencrypted proxy connection. An app showing “connected” doesn’t tell you how the entire connection path is protected.
5. Hysteria2 and WireGuard: check the UDP path first
Hysteria2 is based on QUIC and requires UDP communication to establish a connection. If the current network restricts UDP, Hysteria2 may time out even when a TCP service at the same address is reachable. Hysteria2 settings aren’t interchangeable with those for earlier Hysteria versions. Authentication and obfuscation are separate settings; if the server uses obfuscation, the corresponding parameters must match.
WireGuard uses public-key cryptography to establish a tunnel with a peer. Private Key identifies the local side; enter the corresponding peer’s public key as Public Key. Endpoint is the peer’s external address and port, while the tunnel address belongs to the tunnel’s internal configuration. Using the Endpoint’s public IP as the local tunnel address changes the meaning of the configuration and usually won’t produce the expected routes.
- Check the network connection: Where permitted and practical, compare Wi-Fi and cellular using the same configuration. A failure on only one network is a clue to check for network restrictions, not a definitive diagnosis.
- Check the server port: A TCP port check alone can’t establish whether Hysteria2 or WireGuard will work. Confirm that the server is listening and that UDP traffic is allowed through the firewall.
- Check WireGuard routing: Allowed IPs determine which traffic is sent to the peer and should match the tunnel’s intended use. If the tunnel connects but some addresses remain unreachable, check the allowed range and server forwarding.
- Check MTU last: If small requests work but large ones stall, MTU is one possible area to investigate. Save the original value and adjust it based on testing on the actual connection; don’t rely on a fixed value that supposedly works on every network.
Key takeaway: investigate the network layer first when a connection times out
If Hysteria2 or WireGuard times out, first confirm UDP connectivity and server settings, then check routing. Don’t try to fix the same server configuration by switching to a different Type.
6. Import from Home, then verify with Global Routing
The steps below describe common mobile workflows in Home, Config and Settings. Layouts may differ by platform, so look for the feature with the same name. A subscription provides or updates connection entries; it doesn’t select the right rules configuration for you. After importing, check which entry is selected and the current routing mode.
-
Check the source details
Have the existing service’s protocol type, Address, Port, credentials and additional parameters ready. If you received a subscription URL, make sure it’s still valid and don’t post the full URL publicly if it contains a token.
-
Choose an import method
To add a connection manually, go to Home → tap “+” in the upper-right corner → Type, then select the type that matches the server. To import a subscription, select Subscribe under Type, enter your own subscription URL and save.
-
Check the imported settings
Return to Home and, if needed, pull down to refresh the subscription. Then open the actual entry’s edit screen and check Type, Address, Port and authentication fields. A subscription refresh may overwrite manual changes, so note the original values before testing.
-
Test routing separately
Under Home → Global Routing, Proxy routes traffic through the proxy, Direct connects directly, Config follows the current configuration, and Scene follows the scene settings. You can temporarily select Proxy to rule out rule-based routing as a factor, then restore the original mode.
-
Restore your usual settings
Turn on the connection and visit a specific target you want to test. If it works with Proxy but fails with Config, open Config and check the active file and rules. Then go to Settings → On Demand and review the automatic connection conditions so you don’t mistake automatic reconnection for protocol behavior.
Rules are matched in the order specified in the configuration. DOMAIN-SUFFIX checks domain suffixes, GEOIP uses an IP geolocation database, IP-CIDR matches address ranges, and FINAL handles requests that didn’t match earlier rules. These keywords determine where requests go; they don’t convert VLESS to Trojan or fix an incorrect Password. After testing, restore the original Global Routing and automatic connection settings.
7. Common questions about protocols and imported configurations
Which protocol is always faster?
There’s no fixed ranking independent of network conditions. Keep the test destination, network and time as consistent as possible, then compare connections offered by the same service. Note whether pages load, connections drop and data transfers remain stable. Don’t switch protocols based on a single latency reading.
Can I change VMess Type to VLESS?
Changing Type alone won’t convert the connection. Get the corresponding VLESS configuration from your existing provider, then create or import it from Home → “+” → Type. Even if Address and UUID are the same, the server must support the corresponding protocol, and transport and security settings must also match.
Why can’t I open websites even though the subscription has entries?
First check which entry is selected in Home, then check Global Routing. If Proxy works but Config fails, check the matching rules and policies in the active Config. If both modes fail, check Address, Port, authentication, TLS and network connectivity one by one.
Should I disable validation when I get a TLS certificate error?
Not yet. Check that the device clock and TLS hostname are correct, and make sure you haven’t copied the Address IP directly into the hostname field. Then check the server certificate. If the settings came from a subscription, refresh it and review them again. If the error persists, ask your existing provider to confirm the server configuration.
Do I need to choose a different protocol on a computer or Android device?
Don’t change the protocol just because you’re using a different platform. First use Shadowrocket for that platform, confirm the import format and required features, then check the settings for the same service. Keep network conditions comparable when testing across devices; a connection working on one device doesn’t mean the other has the same network, routing or DNS settings.
8. Configuration checklist
Shadowrocket is a closed-source commercial app. Getting the client and obtaining connection credentials for an existing service are separate matters. This guide covers importing and configuring connections only; it doesn’t provide, sell or recommend proxy servers or subscriptions. For the iPhone / iPad App Store listing, verify the developer is Shadow Launch Technology Limited and the app ID is 932747118. Check the App Store page for system requirements.
- Before configuring: Save your current Config and original connection settings. Confirm the Type provided by the server; don’t guess the protocol from the port or entry name.
- First test: Verify the address, port, authentication and transport settings, then use Global Routing to distinguish connection issues from rule issues.
- After a subscription refresh: Check the selected entry again and see whether manual changes were overwritten. A successful subscription refresh doesn’t mean every entry is usable.
- After testing: Restore the original Config, Global Routing and On Demand settings. Remove sensitive credentials from anything you share.