V2Ray Beginner FAQ: 10 Answers on Clients, Subscriptions, Cores, and Proxy Modes

1. What are V2Ray, a client, and a server node?

The client provides the configuration interface, the core handles traffic, and a node contains the settings for a remote connection. They are different things: installing v2rayN does not give you a server, and importing a node does not mean your browser is already using the proxy.

“V2Ray” can refer to the project itself, but it is also commonly used as a general term for this type of proxy tool. When troubleshooting, avoid saying only “V2Ray isn’t working.” Note the client name, core, proxy mode, and error message so you can pinpoint the relevant settings.

2. Which client should you install on Windows, macOS, Linux, or Android?

Start with v2rayN on desktop, and choose v2rayNG or v2flyNG on Android based on the core your nodes require. Check your operating system and architecture first, then your node requirements. There’s no need to install several clients and try them one by one.

Match the package to your architecture, too: Windows commonly uses x64; macOS packages are available for arm64 and x64; Linux packages may be in deb or rpm format. For Android, arm64 is usually the right place to start. If you’re unsure about your device’s architecture, check the download page for universal package details. Installers and portable archives launch differently, so follow the instructions for the package you chose.

3. Where do subscriptions come from, and how are they different from individual share links?

A subscription URL comes from a server operator or the management panel for a self-hosted service. The client does not automatically create usable remote nodes. A subscription is an endpoint for updating a set of configurations; an individual share link usually represents just one connection.

• Subscription URL: Usually an HTTPS URL that may contain an account token. Add it to a subscription group, then update the group to fetch the node list.
• Share link: Often starts with vmess:// or vless://. Import it from the clipboard instead of entering it in the subscription URL field.
• Configuration file: May be a core configuration or a file exported by a client. Check the supported formats and import method separately; not every JSON file is a subscription.

Getting zero nodes after the first update doesn’t necessarily mean the client is broken. Check the subscription’s expiration and data allowance, then see whether the server returned node data, a login page, an error, or a redirect. A page opening in your browser doesn’t mean the client can parse it as a node.

4. How do you import a subscription into v2rayN or v2rayNG?

On desktop, create a group, update it, then select a node as the active server. The steps below use common v2rayN menu labels. Button locations may vary by release, but look for “Subscription Group Settings,” “Update Subscription,” and “Set as Active Server.”

1. Create a group

   Open “Subscription Group” → “Subscription Group Settings” from the main window. Select “Add” or “+” to create a group, then enter a descriptive name.

2. Paste the URL

   Paste the full subscription URL into the group’s URL field. Check that the beginning, end, and token are intact, then save. Don’t paste an entire chat message with quotation marks.

3. Update the subscription

   Open the “Subscription Group” menu and select “Update All Subscriptions (Direct).” If no nodes are available yet, start with a direct update. If you need a proxy and already have a working node, try updating through the proxy.

4. Choose a server

   Right-click a node in the list and select “Set as Active Server.” Check the logs to confirm the core starts successfully, then configure the system proxy. A node appearing in the list alone isn’t enough.

In v2rayNG, the usual path is “Subscription Group Settings” in the sidebar → “+” in the top-right corner. Enter a name and subscription URL, then save. Return to the main list and select “Update Subscription” from the top-right menu. Choose a node and tap the connect button; the first time you connect, approve Android’s VPN connection request. Menu labels may vary slightly, so use the subscription management option in your installed version.

5. How do you choose between the Xray and V2Fly cores?

Choose a core based on your node’s requirements, not assumptions about speed. VMess and VLESS are protocol names; WebSocket is a transport; TLS and REALITY are security-related settings. These layers can be combined, but not every core supports every combination.

In v2rayN, look for core type settings under “Settings” → “Parameter Settings.” Some versions let you choose a core separately for each protocol. Restart the core after making changes. If the logs report an unknown field or transport type, check the node configuration and core capabilities instead of repeatedly changing the server port. Starting successfully is only the first step; authentication settings and the remote handshake must also be correct.

6. What’s the difference between system proxy and TUN mode?

Start with the system proxy to test browser access. Consider TUN only if an application doesn’t follow the system proxy settings. The system proxy adds the proxy address to your system settings. TUN creates a virtual network interface that can route traffic from more applications, but it also requires attention to permissions, DNS, and local network access.

In v2rayN’s system proxy menu, start with “Set system proxy automatically.” When you’re done, select “Clear system proxy.” Enabling TUN won’t fix an invalid subscription or incorrect password. It changes how traffic reaches the client; it doesn’t replace the node connection itself.

7. When should you use global routing or split tunneling?

For everyday use, start with a suitable routing rule set. For troubleshooting, briefly switch to an explicit proxy route as a comparison. The proxy entry point and routing policy are separate settings: an app first sends traffic to the client, then the core decides whether to send it direct, through the proxy, or block it.

• Direct: The connection bypasses the selected proxy server. Use it for destinations that need a local exit.
• Proxy: Sends traffic through the configured proxy outbound. The selected active node typically applies at this stage.
• Block: Rejects matching connections. This isn’t a slow connection and shouldn’t be mistaken for a disconnected node.
• Rule order: Rules are generally evaluated in order. A broad rule near the top may override more specific rules below it.

Split-tunneling logic (example only; not an importable configuration):
Local network addresses → direct
Domains that need the proxy → proxy
All other connections → use the current route’s default outbound

“Auto-configure” in the system proxy menu doesn’t mean “automatic split tunneling.” Check the active routing preset and the final default outbound, too. If your rules use geosite:cn or geoip:cn, the corresponding rule data must be available. A rule name alone doesn’t provide a complete domain classification.

8. Why won’t a website load even though the node has latency?

A latency result only confirms that one test succeeded; it doesn’t mean the full connection path is working. A successful TCP probe may only mean the remote port is reachable. Loading a website also involves the proxy handshake, authentication, DNS, TLS, and routing. Check what the test measures, not just the number of milliseconds.

Change one setting at a time and record the errors before and after. Subscription update errors occur while fetching configuration, core startup errors occur locally, and remote authentication errors occur during connection. Treating them all the same can break settings that were already working.

9. Which local port should you use, and how do you check for conflicts?

Applications connect to the client’s local listening port, not the remote server’s port. In v2rayN, check the actual listener settings under “Settings” → “Parameter Settings,” then enter them in your app’s proxy settings. The example below uses 127.0.0.1:10808 to demonstrate how to check; this isn’t guaranteed to be the default in every version.

# Windows PowerShell: Check whether the example port is listening for TCP connections
Get-NetTCPConnection -LocalPort 10808 -State Listen

# Linux: List TCP listeners and look for 10808 in the output
ss -lntp

# macOS: Check which process is listening on the example port
lsof -nP -iTCP:10808 -sTCP:LISTEN

If nothing is listening on the port, first check whether the core has started. If the logs show address already in use, find the process occupying the port, close any duplicate client instances, or change the listening port. After changing it, update the app’s manual proxy address too; it won’t update automatically.

10. How do you verify the setup and keep a recoverable configuration?

Use real connection tests, logs, and the results of switching modes to verify your setup—not the “connected” icon alone. After your first successful configuration, note the client, core, listening port, and routing preset. That gives you a baseline for upgrades and node changes.

1. Check startup: After selecting an active server, confirm the core starts normally and the logs don’t repeatedly report a port conflict or configuration parsing error.
2. Check access: Test one destination expected to connect directly and one expected to use the proxy. Check the logs to see which outbound each connection actually uses.
3. Check DNS: If domain names fail but other connections work, check your DNS settings. Don’t judge an HTTPS service solely by connecting directly to its IP address.
4. Check shutdown: If you’re using the system proxy, clear it before closing the client and confirm normal network access returns.
5. Back up your configuration: Use the client’s backup or export feature and store files containing subscriptions or authentication details securely. Remove sensitive information before sharing troubleshooting details.

Keep a recoverable configuration before upgrading. After the upgrade, test with the original node and routing first; don’t change the core, DNS, and TUN settings all at once. If you can’t identify an error, gather six details: operating system, client version, core, protocol combination, reproduction steps, and redacted logs. Then work through the FAQ or complete guide one issue at a time. This narrows the problem down to a specific part of the setup instead of leading to repeated reinstalls.