This VPN beginner’s guide starts with the practical steps: import a configuration using a subscription, choose a connection endpoint from the node list, then use routing rules to decide which requests go through the proxy. The protocol defines how the client and server communicate. Keeping these concepts distinct helps you pinpoint whether an issue lies with the configuration, connection, or rules.
Subscription Links: Importing a Configuration Isn’t Connecting
A subscription link in a client is an address used to retrieve configuration. It is neither a node address nor a client download link. Here, “subscription” describes how configuration is fetched, not a billing period. A client may retrieve a list of nodes, or a configuration containing policy groups, routing rules, and DNS settings—the exact contents depend on the subscription format. A successful import only means the configuration was readable; it doesn’t confirm that any listed node can connect.
- First, follow the client setup guide to confirm which client and configuration format suit your system. Get the matching subscription link from your account panel. Don’t enter it in a form for adding a single server.
- Open the subscription or remote configuration page in your client, paste the link, and update it. Check that the node list appears, then select the configuration; some clients don’t make an imported configuration active automatically.
- Select a node, enable the connection as prompted by the client, then test it with the app you intend to use. If the update fails, first check that the link is complete, still available in your account panel, and accessible to the client. If the update succeeds but the connection fails, check the node and your local network separately.
Nodes and Routes: A Label Doesn’t Show the Full Path
A node is a connection configuration you can select in a client, usually containing a server address, port, and protocol-specific settings. A route is the path your traffic takes. Between the first endpoint your device connects to and the exit seen by the destination website, traffic may pass through relays. A region label in the list usually describes the exit location or service location; it doesn’t tell you where the first endpoint is. Two nodes labeled with the same region may take different routes.
| Label | What the Route Means | What to Check |
|---|---|---|
| Direct node | The client connects straight to the node endpoint, without an additional relay endpoint configured by the provider for that route. | The quality of your local network connection to the endpoint affects performance. Here, “direct” is different from DIRECT in routing rules, which means a request doesn’t use the proxy. |
| Relayed node | The client first connects to a relay endpoint, and the server forwards traffic to the next leg of the route. | A relay isn’t the same as a dedicated line. An accessible first endpoint doesn’t guarantee a stable connection on the rest of the route. |
| IEPL dedicated line | IEPL stands for International Ethernet Private Line. The label usually indicates that a segment of the service route uses a dedicated line. | A dedicated-line label doesn’t mean the entire path from your device to the website is exclusive, nor does it indicate whether the exit is compatible with your needs. Judge it by how your apps perform. |
When choosing a node, first check whether its exit region meets your work or app requirements. Then compare performance on your usual network and at the times you typically connect. A latency reading in the client may reflect just one probe; it doesn’t indicate sustained throughput or whether video calls will be stable. If a route loads webpages quickly but often drops file uploads, test it with an upload task rather than relying on the latency ranking.
Choosing a Protocol: Check Support, Then Your Network
Subscription clients often list several proxy protocols together. These aren’t the same as the VPN configurations built into your operating system: being able to add a VPN in your system settings, or create a virtual network interface in a client, doesn’t mean it supports every protocol in the list. The server determines which protocol a node uses. You can’t make a node work by simply changing its protocol label.
- Shadowsocks: An encrypted proxy protocol. When importing a node, check that your client supports its encryption method. If the configuration requires a plugin, check plugin compatibility too.
- VMess and VLESS: These configurations aren’t interchangeable. VLESS doesn’t provide transport encryption on its own, so check the transport and security layers configured for the node. If the connection fails after import, don’t try to fix it by disabling certificate verification.
- Trojan: Typically establishes connections over TLS. The server name, certificate verification, and server settings must match; copying the port and password alone may not be enough for a complete configuration.
- Hysteria2 and TUIC: Both use QUIC over UDP, but they aren’t compatible with each other. If your network restricts this UDP traffic, a correct configuration may still fail to connect. If available, test a TCP-based node provided by the service.
When comparing protocols, first make sure your client supports the server configuration, then keep test conditions as similar as possible. Switching protocols often also means switching servers or exits, so a speed improvement can’t automatically be credited to the protocol. For beginners, a reliable default is a node that imports as configured and consistently handles the tasks you need on your usual network.
Global Mode and Rule-Based Routing: What Happens After Traffic Is Captured
Global mode usually sends requests already captured by the client through the selected outbound. Rule-based mode decides whether to proxy, connect directly, or reject requests based on domains, destination addresses, or other conditions supported by the client. The key is “already captured”: if you enable only the system proxy, apps that ignore system proxy settings may never reach the client. Clients with a virtual network interface mode can capture more traffic, but routing scope, permissions, and exclusions still matter.
For everyday use, start with rule-based mode. For example, access your company intranet as required by your organization, and route other permitted proxy traffic according to the applicable rules. If a website won’t load in rule-based mode, you can briefly switch to global mode on a network where that’s allowed and compare. If it works globally, switch back and check which rule matched the request, which outbound the policy group selected, and whether sign-in pages or images took a different route. Many clients apply rules in order, so a broad rule matched earlier may override a more specific one. Use the client’s match log to confirm what happened.
Platform differences can also affect the result. On desktop, distinguish the system proxy from a virtual network interface. Mobile clients may need system permission to capture traffic, and per-app routing depends on the operating system and client. A browser test can’t stand in for testing a desktop program or mobile app.
Check Your Exit IP and DNS After Connecting
A client showing “connected” doesn’t mean the app you’re testing is using the expected exit. DNS translates domain names into addresses and has its own request path. A DNS leak is an issue when a service requires domain lookups to use a specified resolver path, but a query bypasses it—for example, by going to a resolver on your local network. In rule-based mode, a domain explicitly configured for direct access may use local DNS; that alone isn’t necessarily a leak.
- ✅ First, note your exit IP before connecting. Then open My IP in the same browser and compare. In rule-based mode, confirm which outbound the check page itself should use so you don’t mistake an expected direct result for a problem.
- ✅ Open the app you actually want to use and check the client’s connection log, if available. If there’s no trace of the request, check the traffic capture scope. If the request appears but uses the wrong outbound, check the routing rules and policy group.
- ✅ Check your system DNS, client DNS, and browser secure DNS settings. The browser may use its own encrypted DNS or a cached result. Check the resolution path for a new query; an exit IP change alone doesn’t prove DNS is being handled as expected.
- ❌ Don’t change networks, nodes, protocols, and rules all at once while troubleshooting. Keep other conditions fixed and change one thing at a time so you can identify what actually fixed the issue.