Quick verdict: Choose by platform, then by traffic coverage

If you mainly use existing subscriptions and want a consistent workflow across Android, iOS, and desktop, start with Clash Plus. Test importing your subscription, switching policies, and refreshing it before deciding whether to move your other devices. For per-app traffic controls on Android, compare Clash Meta for Android. For frequent rule debugging, connection monitoring, or switching between system proxy and TUN on desktop, compare Clash Verge Rev.

Client Platform and core version Traffic handling Subscriptions and interface
Clash Plus Android, iOS, Windows, and macOS; check subscription formats and required fields on each platform On mobile, check system VPN permissions; on desktop, confirm the traffic-handling options in your installed build A consistent starting point across devices; first locate your commonly used policy groups and manual refresh controls
Clash Meta for Android Android; based on Clash Meta (mihomo)—check the version bundled with the APK Routes traffic through Android VpnService; test app allowlists or exclusions if you need per-app controls Useful for managing configurations and proxy groups separately while keeping mobile settings independent
Clash Verge Rev Windows, macOS, and Linux; based on mihomo—check the core version actually running System proxy and TUN are separate options; TUN also depends on system permissions and service status Useful for checking rule matches across subscriptions, proxy groups, and the connection list
FlClashs Android, Windows, macOS, and Linux; check the core details for the specific build Check Android VPN permissions and the traffic-handling options in desktop builds separately An alternative for phones and desktops; compare configuration switching and update controls

On iOS, you’ll need an iOS client—not an Android APK or a desktop core ported to iOS. To see whether a subscription is compatible, check that the iOS client can parse its configuration, show the expected policy groups, and establish a connection. Desktop toggles and editable fields may not have direct equivalents on iOS.

Core version: the client version alone doesn’t tell you whether a configuration is compatible

The client handles the interface, subscriptions, and system integration; the core parses configurations, matches rules, and establishes connections. Two clients that both list mihomo may bundle different core versions, and Android and desktop builds of the same client may differ too. If you’re still using Clash for Windows 0.20.39 on an older device, don’t use that client version to infer support for current mihomo configurations. Test your configuration before replacing the old setup.

Check these three areas with your actual configuration

  1. Node types. Find a node used in your current configuration. For example, if it uses type: vless with reality-opts, test an actual connection through that node. Seeing a node in the list doesn’t mean all its connection parameters were handled correctly.
  2. Policy groups and rule providers. If your configuration uses proxy-providers or rule-providers, check that the provider data loads successfully. Then inspect the rule and outbound used by a test connection. A successful subscription import alone doesn’t confirm that remote rule providers are working.
  3. Device-specific settings. Mobile clients may not apply desktop settings such as tun, listening ports, and DNS exactly as configured. Note the core version in use; if it isn’t shown in the interface, check the build’s release notes or startup logs, then use configuration errors to identify the specific field.

TUN, system proxy, and mobile VPN: what’s the difference?

Desktop: browser works, but terminal traffic may not

A system proxy mainly affects apps that read the system proxy settings. If a website loads in your browser, that doesn’t mean terminal commands, apps with their own network stack, or local network tools use the same route. For browser-only use, try the system proxy first. Consider TUN if an app doesn’t use the system proxy and has no proxy setting of its own. After enabling TUN, check local network access and DNS resolution too, so routing issues aren’t mistaken for a faulty proxy node.

For example, if you’ve confirmed the client exposes an HTTP or mixed port on your device at 7890, you can explicitly set the proxy in a desktop terminal. This example uses example.com as the test destination. In Windows PowerShell, use curl.exe instead of curl.

curl --proxy http://127.0.0.1:7890 --connect-timeout 10 -I https://example.com

7890 is an example, not a default shared by all clients. A successful command only means the request went through the local proxy port; if a rule routes the domain to DIRECT, it still bypasses the proxy node. To confirm that a proxy node handled the request, check the client’s connection log at the same time. To test TUN, remove the explicit proxy arguments, make sure the terminal hasn’t inherited proxy environment variables, and check whether the connection appears in the client.

Android and iOS: verify the tunnel, then test your apps

On Android, granting VpnService permission lets the client create a local VPN interface; routing is still determined by your rules. If the client has an app exclusion list, test one app on the list and one that isn’t. The VPN icon only shows that a tunnel is active—it doesn’t confirm that every app’s traffic is being captured. If another VPN is already connected, check which one is active during your test.

On iOS, if a system prompt asks to add a VPN configuration at startup, first confirm that the newly installed client initiated the request. After granting permission, test two destinations with known routing outcomes—for example, one configured for direct access and one configured to use a proxy. Don’t judge whether an iOS client can handle traffic by looking for a separate TUN toggle like the one in desktop clients.

Subscription management: test import, activation, and refresh separately

A subscription URL is an entry point for configuration updates; the imported configuration is what the client can actually use. The same URL may return a complete Clash YAML file or a node-only subscription, depending on the provider. If an older client shows policy groups but a newer one shows only nodes, first check whether both received the same format, then confirm that the new client loaded the intended configuration.

  1. Confirm the configuration is active. After importing, check the name of the active configuration. If you keep a temporary test configuration, give it a distinct name so you don’t keep using the old one after a successful refresh.
  2. Check whether your policy selection persists. Select a node in a manual selection group, refresh once, and check whether your selection remains. If the provider removes a node or renames a policy group, the client may revert to its default selection.
  3. Keep provider updates separate from local changes. Changes made directly to a subscription-generated file may be overwritten on refresh. If the client offers an override or merge feature, put local DNS and rule additions there. Otherwise, save a separate test configuration instead of treating the original subscription as a file for ongoing manual edits.
  4. Check what happens when an update fails. During a manual refresh, note the status message, timestamp, and active configuration. On mobile, also test after switching between Wi-Fi and cellular data. An option to set a refresh interval doesn’t guarantee the system will run updates in the background on schedule.

Subscription URLs may contain access tokens. Import them only on your own devices during a migration. When reporting an issue, include the configuration name, error field, and time—but don’t share the full URL in screenshots, support tickets, or public discussions.

Interface trade-offs and migration checks

For a first-time setup, focus on three things: can you quickly identify the active configuration, find the policy group in use, and understand why a connection failed? Clash Plus is a good place to start when checking the basics of one subscription across your usual devices. If you often change per-app settings on your phone, focus on Clash Meta for Android. If you check rules and outbound connections every day on desktop, give Clash Verge Rev’s connection view a closer look. You can use different clients on different devices—there’s no need to sacrifice essential traffic controls just to keep the interface consistent.

Test a migration with two requests

  1. Keep the old configuration, then import the same subscription into the new client. Confirm the active configuration and policy groups, then manually select a node you know works. Don’t choose a node based only on the lowest latency-test result.
  2. In rule mode, visit one domain expected to go direct and one expected to use a proxy. If connection logs are available, check the rule and outbound for both requests. If you can only check the exit IP, test the same destination separately each time and first confirm its routing policy.
  3. Include the apps you actually use in your tests. On mobile, switch networks and test again after locking the screen. On desktop, retest with the apps whose traffic you need to capture. Finally, refresh the subscription manually and repeat a proxy request.

If only one platform fails, first compare the active configuration, traffic-handling mode, and core errors—don’t change the subscription, DNS, and policy groups all at once. For system settings, see the user manual. To compare multiple desktop clients, see client comparisons.