At a glance

If you already have connection details, a share link, or a subscription from your provider, start with the protocol specified by the server, then check the authentication, transport, and TLS settings. This guide uses Add Server on iPhone / iPad to explain what each field corresponds to and how to troubleshoot whether the issue is an incorrect setting, a network problem, or traffic that is not being routed through the proxy.

Protocol comparison: What to enter in Add Server

A domain can provide different services on different ports, and the same port number can use different transports. The address and port tell the client where to connect; Type, authentication details, and transport settings determine whether the two ends can establish a session. The table lists details to confirm with your provider; field labels are not standardized across platforms.

How each protocol works and what to check when configuring it manually
ProtocolHow it worksDetails to check
ShadowsocksWraps proxy traffic using an agreed encryption method.Server address, port, encryption method, and password or key. If the configuration includes a plugin, also check its type and parameters.
VMessAuthenticates with a user ID; match the proxy protocol and its underlying transport separately.Address, port, UUID, and transport type. For WebSocket, check Host and Path; if TLS is enabled, also check the server name.
VLESSAuthenticates with a user ID; the commonly used encryption=none setting relies on the outer security transport for protection.Address, port, UUID, transport, and security settings; also check flow if specified in the original configuration.
TrojanTypically authenticates with a password and forwards traffic over a TLS connection.Address, port, password, and TLS server name. Use the original settings for any additional transport parameters specified by the server.
HTTPForwards requests through an HTTP proxy; requests to HTTPS destinations typically use CONNECT to establish a tunnel.Proxy address and port; enter a username and password only if the proxy requires authentication. An HTTPS destination does not mean the connection to the proxy also uses TLS.
SOCKS5The proxy can forward TCP; UDP availability depends on whether both the client and server support UDP forwarding.Proxy address, port, and any authentication details required by the server. A webpage loading successfully does not confirm that UDP is available.
WireGuardUses public and private keys to identify the peer and carries IP packets over UDP.Your private key and tunnel address, peer public key, Endpoint, and AllowedIPs. Keep the pre-shared key too if the original configuration uses one.
Hysteria2Carries proxy traffic over QUIC, which is based on UDP.Address, UDP port, authentication value, and TLS server name. Enter an obfuscation type and key only if obfuscation is enabled on the server.

Ports, transport, and TLS: similar names do not mean identical settings

443/TCP
The default TCP port for HTTPS; this does not tell you which port the proxy uses.
443/UDP
Hysteria2 can use this port; its UDP port is independent of the TCP port with the same number.
Two key roles
For WireGuard, distinguish your private key from the peer's public key.

Bottom line: Check which transport the server is listening on first

If your provider specifies 443/UDP but a test shows only 443/TCP is reachable, that does not rule out a UDP network issue. Keep the original settings and test again on another network that allows UDP. Do not switch to a different protocol just because it uses the same port number.

VMess or VLESS configurations using WebSocket may include a connection address, an HTTP Host, and a TLS server name (SNI). The connection address identifies the host to connect to, Host identifies the site at the HTTP layer, and SNI tells the TLS layer which site and certificate to use. These values may be the same or different. If the certificate name does not match, first check that SNI matches a domain covered by the certificate. Disabling certificate verification is not a fix.

In WireGuard, Endpoint is the peer's externally reachable address and port, not your tunnel address. AllowedIPs specifies the IP prefixes that can be reached through the peer; it is not a port allowlist. If a handshake succeeds but you cannot reach a destination inside the tunnel, check whether its address falls within the relevant prefix, then check the peer's routes. Do not share your private key or full configuration while troubleshooting.

Add a server manually and test the connection

Manual setup is for when you have the complete connection details for a single server. If you already have a subscription, go to Home → + → Add Server → Type, select Subscribe, enter your subscription URL, and save. A subscription URL retrieves a list of servers; it is not the address of a single server, so do not split it into Address and Port.

  1. Choose a protocol

    Go to Home → + → Add Server → Type and choose the protocol specified in the original details. Both VMess and VLESS can use a UUID, so the UUID alone does not tell you which one to select.

  2. Enter the server address and port

    Enter the server hostname or IP address and port in Address and Port. Do not paste a complete share link into Address. For WireGuard, the peer's connection address and port belong in Endpoint, not in the tunnel address fields.

  3. Check authentication and transport settings

    Use the original details to enter the password, UUID, or key, then check the encryption method, transport type, and TLS settings. If a WebSocket Path is specified, such as /edge, keep the leading slash; do not include the Path as part of the address.

  4. Save and test the connection

    Save, return to Home, select the server you just added, and turn on the connection switch. First, use Connectivity Test to record the result for that server, then try a destination known to be reachable. A successful test means only that the test request succeeded; it does not prove that every site is reachable or that UDP forwarding works.

  5. Restore your previous routing mode

    To check whether the issue is caused by rules, temporarily select Home → Global Routing → Proxy in a test environment where proxy use is permitted, then test the same destination again. Restore your previous mode afterward. In Config mode, the active configuration still determines how traffic is routed.

The protocol connects, but requests are not going through the proxy

Add Server defines how to connect to the server; Global Routing determines how requests are routed. Direct connects directly, Proxy forwards traffic through the proxy, and Config routes traffic according to the active configuration. Scene selects a routing mode based on the configured scene. Turning on the connection does not mean every request will use the protocol you just added.

[Rule]
DOMAIN-SUFFIX,example.com,PROXY
IP-CIDR,192.168.0.0/16,DIRECT
FINAL,DIRECT

This configuration snippet illustrates rule matching order; it is not a template to replace your existing configuration. In Config mode, a request to the recognized domain www.example.com can match the first rule and use PROXY; a direct request to 192.168.1.1 can match the second rule and use DIRECT. Requests that match none of the earlier rules connect directly under FINAL,DIRECT. To troubleshoot, check which file is active in Config, then look from top to bottom for the first rule matching the destination. The mere presence of PROXY somewhere in the file does not mean the request will use it.

Common issues: timeouts, certificate errors, and subscription updates

Does Hysteria2 time out on Wi-Fi but connect over cellular?

Test the same server and destination on both networks without changing the port or TLS settings. If the issue occurs on only one network, first check whether that network allows access to the server's specified UDP port. Being able to visit HTTPS sites over that network does not prove that the UDP port with the same number is reachable.

Why does Trojan report a certificate name mismatch when configured with an IP address?

Return to Home, edit the server, and check that the TLS server name is the domain provided by your service provider and covered by the certificate. Even if the connection address must be an IP address, the certificate may be valid only for a domain. If the name is correct and the error persists, check the device time and the server certificate status.

Why does a UUID work with VLESS but not with VMess?

A UUID is an authentication parameter, not a protocol identifier. Go to Add Server → Type and select VLESS as specified in the original configuration, then check the transport, security settings, and any server-specified flow one by one. Keeping the UUID and changing Type does not convert a configuration to another protocol.

One server in a subscription times out. Should I delete and re-add the subscription?

First, test another existing server under the same subscription in Home. If only one server fails, note its name and the network in use, then check whether the subscription has been updated with the current details from your provider. If the entire list is empty, check that Type is set to Subscribe and that the URL was copied in full and is still valid. When reporting the issue, hide any tokens and authentication details in the subscription URL.

Get Shadowrocket and browse setup guides

Get Shadowrocket for iPhone / iPad on the App Store for a one-time payment of $2.99. The developer is Shadow Launch Technology Limited, and the app ID is 932747118. Check the App Store page for system requirements. The Mac (Apple Silicon / Intel), Windows, Android, and Linux versions are free; download packages from this site's download page.