If you already have a subscription, import its nodes as provided—don't edit an existing node just to switch protocols. If both protocols are available, first make sure your client core supports the required transport and security settings, then compare connection performance under the same conditions. The sections below explain how to distinguish these settings and what to check first when a connection fails.
Encryption in VMess and VLESS
Node details can make it easy to confuse the protocol, transport, and transport security. VMess and VLESS are proxy protocols that carry user identity and destination information. WebSocket and similar options are transports; TLS and REALITY are security settings. “VLESS + TLS” doesn't mean choosing between two protocols—it describes different layers of the same connection.
VLESS
RecommendedIn common configurations, the protocol encryption field is none; TLS or REALITY protects the connection. Try this setup only when the server provides the matching configuration and your client core supports it.
Best for: subscriptions that offer both types of nodes, when the VLESS node has complete transport and security settings.
VMess
The protocol defines its own authentication and data encryption, and authentication also depends on the time. Nodes can also use TLS; outer-layer TLS and VMess encryption are separate settings.
Best for: subscriptions that only offer VMess, or an existing configuration that works reliably.
encryption=none doesn't mean the entire connection is unencrypted—check whether transport security is enabled too. Conversely, if both the protocol and transport security fields are set to none, a commonly used HTTPS port doesn't mean the connection is protected by TLS. Check the actual configuration; don't guess from a node label or port number.
How much overhead does the transport add?
Take a common VLESS request format as an example. With no additional information, the version field uses 1 byte, the user ID 16 bytes, and the additional-information length field 1 byte. For a TCP command and an IPv4 destination, the command, destination port, address type, and address use 1, 2, 1, and 4 bytes respectively, bringing the request header to 26 bytes total.
These byte counts describe only the request header—not the response header, application data, or outer encapsulation—and aren't fixed overhead for every network packet. When the destination is a domain name, the request header also includes the domain length and name. VMess overhead depends on the specific header format and data encryption and framing options. You can't infer it from speed-test results for two arbitrary subscription nodes.
Compare under the same conditions before switching
If the server offers both configurations on the same region, route, and transport, alternate between them on the same device and network, testing the same destination. Record initial load time, sustained download performance, and failure rate. If the nodes use different routes, you can only determine which works better on your current network—not attribute the result to the protocol.
Choosing a node on desktop and Android
After importing the same subscription into v2rayN or v2rayNG, test one node at a time rather than changing the entire group. A client may display every subscription entry without the active core being able to run every combination. For VLESS nodes in particular, check the transport type, security settings, and any flow-control requirements—not just the protocol name.
Check the same subscription on both devices
Desktop: v2rayN
- Open “Subscription Group,” select “Update subscription,” and confirm the group now shows the new nodes.
- Double-click a node to inspect its configuration. Check the address, port, transport type, and transport security. If the subscription uses REALITY, make sure the active core supports it.
- Set the node as the active server, check the current mode under “System Proxy” in the tray menu, then test it in a browser.
Android: v2rayNG
- After adding the subscription, open the top-right menu on the main screen, select “Update subscription,” then tap the node you want to test.
- Use the edit option next to a node to check its protocol and security settings. If you're only inspecting it, exit without saving. v2rayNG uses the Xray core, but supported features depend on the core version bundled with the client.
- Tap the connect button. On first use, approve the VPN connection request when prompted by Android, then test with an app that needs network access.
Menu labels may vary slightly by version. If the two clients behave differently, first make sure you've selected the same node, then check the core and which traffic is routed through the proxy.
If you use v2flyNG on Android, don't choose nodes based on v2rayNG's Xray feature list. v2flyNG uses the v2fly core, so choose a protocol combination supported by that core and your client version. If the subscription says a node requires an Xray feature, use a client that supports it rather than deleting unfamiliar fields and trying again.
Can't connect after importing? Start here
- First, make sure the subscription updated successfully. A failed subscription fetch, a missing group in the subscription, and a node that won't connect are three different problems. If the list hasn't refreshed, check the subscription group and update status. If the list looks right, try another known-working node in the same group for comparison.
- Next, distinguish a handshake failure from traffic not being routed. Check the client logs after starting the connection. For authentication or handshake errors, check the node settings first. If the client is running but your test app has no traffic, check the system proxy and whether the app respects it on desktop; on Android, check that the VPN is connected and the app is included in per-app proxy settings.
- Finally, check protocol-specific settings. For VMess, check the device time and user ID first. For VLESS, make sure flow control and transport security match the subscription exactly. If you're using WebSocket, check the path and host settings too. Logs can help identify where the connection fails, but remove subscription URLs, node addresses, and user IDs before sharing them.
VMess connected yesterday, but now reports an authentication failure. Why?
Open your device's date and time settings, turn on automatic time, and let it sync before retrying the same node. Changing the displayed time zone doesn't correct the system clock. If the time is correct and the connection still fails, update the subscription and check whether the node has been replaced.
Can I change a VMess node to VLESS directly?
Changing the protocol in the client won't migrate a node. Even if the address, port, and user ID look the same, the server must be configured with a VLESS inbound. Import a separate VLESS entry provided by the server, and keep the original VMess configuration for comparison.
The address and port are correct. Why is the TLS handshake still failing?
Compare the server name and transport security type in the node details with the original subscription settings. The connection address and the server name used for the handshake may differ. Don't disable certificate verification to suppress the error. If the server name and certificate don't match, ask the configuration provider to check them.
The latency test succeeded. Why won't my app load?
First confirm that the tested node is the active connection, then test again in the app you actually need. A single latency result only shows that one probe received a response; it doesn't prove that every app used the same proxy route. Check the system proxy or per-app proxy settings, then use the logs to decide whether to investigate DNS or routing.