A 10-Second Check That's Faster Than Going Through the List
Before trying each of the 5 items below, there's a faster shortcut: check whether the client's interface shows "Disconnected," "Connecting," or "Connected." If it never even reaches "Connected," the problem is almost certainly No.1 or No.2; if it shows "Connected" but pages still won't load, the problem is more likely No.3 through No.5. This one check saves you half of the unnecessary attempts, instead of trying all 5 items from top to bottom.
No.1: The Node Itself Is Dead (Most Common)
This is always the first thing to suspect when troubleshooting a connection issue. Free or temporary nodes have a relatively high failure rate — even if one worked yesterday, it may stop working today because of a change on the provider's end. The fix is simple: go back to the server list and run a speed test on the current node; if it times out or shows abnormal latency, just switch to a node that tests fine — no need to dig into the exact cause.
No.2: Routing Mode Got Switched to Direct by Accident
The second most common cause is an incorrect routing mode setting — most typically, the mode got switched to "Direct," so no traffic goes through a node at all, even though the client shows "Connected." It's effectively the same as not being connected. To check, open the routing mode settings and confirm it's set to PAC smart routing or Global proxy, not Direct.
No.3: PAC Smart Routing Misclassified a Site as "Should Connect Directly"
If you're seeing "some sites load, others don't," it's usually not a connection problem at all — it's the built-in routing rules classifying that particular site as "should connect directly." You can temporarily switch to Global Proxy mode to check: if the site loads after switching, the routing rules did misclassify it, and you can manually add that site to the "use node" list in Rule mode.
No.4: The System or Security Software Blocked the Client's Network Access
In rare cases, a system firewall or security software will re-block the client's network access after an update or reinstall, so it appears "Connected" on the surface while data isn't actually being forwarded. Check whether the client is allowed to access the network in your firewall settings, or add it to the trusted list in your security software and restart the client to try again.
No.5: Your Network Environment Changed
For example, switching from your office Wi-Fi to a home network, or from a wired connection to mobile data — this kind of network change can occasionally interrupt the connection, and it has nothing to do with the client or the node; it's completely normal. When this happens, clicking connect again usually restores things — no reconfiguration needed.
How We Ranked These, and How to Prevent Them
This ranking isn't a guess — it's based on real statistics from the reports we've received. Dead nodes and incorrect routing mode settings together account for the vast majority of reports, and both have simple fixes that fall into the "just check and confirm" category. Ironically, the thing many beginners initially worry about — "is there a bug in the software itself" — turns out to be a rare cause in practice. Connection issues actually caused by a software defect are very uncommon, so it's worth ruling out the more common possibilities first whenever you run into a problem.
Once you know how the causes break down, preventing them is actually easier than troubleshooting them after the fact. Try building a habit: check the node's latency before each connection, and proactively switch nodes if it looks noticeably high, rather than waiting until you can't connect at all; and set your subscription to auto-update (with a 12–24 hour interval) to reduce how often you need to maintain the node list manually. These two small habits prevent most of the issues above from happening in the first place, rather than fixing them after they occur, and save you a fair amount of troubleshooting time in the long run.
Still Not Fixed After Checking All 5?
In this case, regardless of which item matched your situation, it's worth doing one general step first: keep only one node that tests well, switch the routing mode to Global Proxy, and test a site you visit regularly. This rules out multiple variables interfering at once — if it loads fine in this isolated test, gradually switch your settings back to normal (like PAC routing) one at a time, which makes it much easier to pin down exactly which step is causing the problem, rather than changing several settings at once and having no idea which one actually mattered.
If it's still not fixed after this step, your situation is likely more unusual — we recommend checking the more detailed troubleshooting section in the full guide and working through it by "symptom → cause → fix" for further diagnosis. Most troubleshooting that doesn't get anywhere happens because a step got skipped and a cause was ruled out prematurely — going through the full sequence in order is usually faster than guessing, and less likely to miss the actual cause. After all, the root cause of the vast majority of connection issues is quite mundane, and rarely requires any deep technical wizardry to fix.