· Getting Started · 8 min read

10 Clash FAQs for Beginners: Configs, Proxy Modes, Ports, and TUN Explained

Answers to 10 common Clash setup questions, from subscriptions and config files to proxy modes, mixed ports, and when to enable TUN.

1. What’s the difference between the Clash client and the Mihomo core?

The Clash client provides the interface; the core handles local ports, matches rules, and establishes outbound connections. For example, Clash Verge Rev uses the Mihomo core. In older guides, Clash Meta usually refers to Mihomo’s predecessor—not a second client you need to install.

If a config option isn’t recognized, check the client version and core version separately. The client version affects menus and import steps; the core version determines supported config fields, rule types, and protocols. Don’t assume a config will load just because the app name looks right.

2. Are a subscription link and a YAML config file the same thing?

No. A subscription link is an address that provides content; a YAML config file is a settings document read by the core. A remote link may return a complete Clash-compatible config, a list of nodes, or a format intended for another client. Being able to open a link in your browser doesn’t mean its contents can be loaded directly.

  • Remote subscription: Add a remote config in your Clash client. After importing it, select the config and confirm that the core loads it successfully. Updating it fetches the latest content; it doesn’t switch the active config.
  • Local file: Import an existing .yaml or .yml file, then set it as the active config. The file extension alone doesn’t guarantee that its contents are valid.
  • Node list: A list of nodes may not include policy groups or routing rules. Check whether the provider offers a complete config compatible with Clash or Mihomo.

3. What do nodes, policy groups, and rules each do?

Rules send traffic to an outbound or policy group; policy groups choose which option to use; nodes define how to connect to a remote proxy. In rule mode, for example, a domain may match a rule that points to the “Manual Select” group. The connection then uses the node currently selected in that group. If a rule points directly to DIRECT, switching nodes won’t change how that connection is routed.

When troubleshooting, check the matched rule in the connection details, then check the policy group and actual outbound. Rules are generally evaluated from top to bottom, so put specific domain rules before broader rules that could override them. An unconditional MATCH rule usually goes last to catch any connections that haven’t matched earlier rules.

4. When should you use Rule, Global, or Direct mode?

Use Rule mode for everyday traffic. Global mode is useful for brief comparison tests, while Direct mode helps determine whether a problem is related to proxy routing. These modes only affect traffic that reaches the core; switching modes won’t automatically route apps that aren’t connected to it.

ModeHow it handles trafficBest for
RuleRoutes traffic to an outbound or policy group based on rulesEveryday traffic routing
GlobalUses the current outbound selected in the Global policy group, bypassing normal routing rulesChecking whether a rule selected the wrong outbound
DirectForwards connected traffic directly through the coreChecking whether a proxy node is affecting the connection

If the same address works in Global mode, switch back to Rule mode and check which rule matched. There’s no need to reimport your subscription. Direct mode doesn’t necessarily turn off the system proxy: apps may still send requests to a local port, while the core routes them directly.

5. What is a mixed port, and why is 7890 so common?

The mixed-port setting lets a single local TCP port accept both HTTP proxy and SOCKS5 requests. 7890 is a common example, not a fixed port for every Clash client or the port of a remote node.

When configuring an app manually on the same Mac, the address is usually 127.0.0.1. Use the actual port shown in your Clash client. If the config specifies separate port and socks-port values, use the one that matches the proxy protocol selected in the app. Don’t confuse either with the controller port.

lsof -nP -iTCP:7890 -sTCP:LISTEN

Replace 7890 in the command with the actual port. The result can show which process is listening, but a listening port doesn’t mean the proxy node is reachable. If another process is using the port, identify it first, then decide whether to close a duplicate Clash client or change the port and update the app settings.

6. Does enabling the macOS system proxy route traffic from every app?

No. The system proxy is a network setting that apps can choose to use. Browsers usually support it, but command-line tools, games, and apps with their own network settings may work differently. A working webpage in your browser only confirms that the browser’s request succeeded—not that a terminal command used the same route.

  1. Go to System Settings → Network → your current network service (such as Wi‑Fi) → Details → Proxies. If you switch to a wired connection, check the active Ethernet service rather than relying on old Wi‑Fi settings.
  2. Check which proxy types are enabled, along with the server address and port. For a proxy pointing to the local Clash client, the address is usually 127.0.0.1, and the port should match the corresponding listener.
  3. Make a new request and check the connection log in your Clash client. If browser requests appear but requests from the target app don’t, first check whether that app supports the system proxy or manual proxy settings.

7. How can you test the proxy port from Terminal?

Use curl to specify a proxy for a single request. This is easier to troubleshoot than editing your shell startup files. The examples below assume a mixed port of 7890; check the port shown in your Clash client before running them. They test the HTTP proxy and SOCKS5 endpoint separately, so you don’t need to run both.

curl -v --max-time 10 --noproxy "" --proxy http://127.0.0.1:7890 https://example.com/
curl -v --max-time 10 --noproxy "" --proxy socks5h://127.0.0.1:7890 https://example.com/

The first command creates a tunnel to an HTTPS destination through the HTTP proxy. The second sends the destination hostname to the SOCKS5 proxy for resolution. --noproxy "" makes sure the bypass list doesn’t affect this test. If the local port refuses the connection, check that it’s listening. If the port accepts the connection but the request fails, check the connection log, matched rule, and outbound. One successful request doesn’t mean other command-line tools will automatically use the same settings.

8. When should you enable TUN?

Consider TUN when the target app doesn’t use the system proxy and has no proxy settings of its own. TUN routes traffic through a virtual network interface, making it better suited to apps that don’t support explicit proxies. Which traffic is captured still depends on routing and exclusions. UDP forwarding also depends on the core configuration and outbound support.

  1. First, test the browser with the system proxy. Then confirm that the target app’s new connections don’t appear in the connection log.
  2. Note the current toggle state. Follow the Clash client’s prompts to grant any system permissions required for TUN, and change only one setting at a time.
  3. After enabling TUN, test the target app and then check your work VPN, local network printer, or file sharing. If anything conflicts, turn off TUN and check routing and DNS one at a time.

9. A domain won’t load—is it DNS or the proxy node?

Start by identifying where the failure occurs. DNS resolution failure, an unreachable proxy port, and a proxy outbound timeout are different problems. A browser’s “can’t reach this site” page alone won’t tell you which one it is.

  • DNS resolution failure: Note the domain shown in the error. It could be the destination website’s domain or the node server’s domain. The system, browser, and Mihomo core may use different DNS resolution paths, so a successful lookup in a system tool doesn’t guarantee the core can resolve it.
  • Outbound timeout: If the log already shows a matched rule and selected node, try another available node in the same policy group and retest the same address. Don’t change DNS at the same time.
  • Only fails with TUN enabled: Switch back to your previous connection method for comparison. If the issue goes away when TUN is off, check routing, DNS interception, and conflicts with other VPNs first. There’s no need to replace the entire config right away.

10. What’s the quickest way to verify a first-time setup?

Check each step in order: config loads → app traffic reaches the core → a rule matches → the outbound connects. Start with one browser and a known-good HTTPS address to keep variables to a minimum.

  1. Import and select a compatible config. Confirm the core reports no errors and that policy groups appear with options to choose from.
  2. Select Rule mode and enable the system proxy, leaving TUN off for now. Note the actual listening port and the current selection in the policy group.
  3. Make a new request in your browser. Confirm the destination domain appears in the connection log, then check that the matched rule, policy group, and outbound are as expected.
  4. Next, test Terminal or another app. If there’s no log entry, check how the app connects. If there is an entry but it uses the wrong outbound, check the rules. If the outbound is correct but the connection still fails, check DNS or the node.

Change one setting at a time and retest the same destination. When you’re done troubleshooting, switch back from temporary Global mode to Rule mode. If you don’t need TUN, stick with the simpler connection method.

Download Clash