Shadowrocket Features & SettingsGlobal Routing, DNS, and On Demand
Review settings in this order: routing mode → rule policies → servers → automatic connections → usage and diagnostics. The steps below use the iPhone / iPad interface. Keep one working configuration, then adjust one setting at a time to avoid changing multiple variables at once.
Availability: iPhone / iPad: $2.99 one-time purchase on the App Store. Mac (Apple silicon and Intel), Windows, Android, and Linux versions are free. See the download guide for your device.
Configroutes traffic according to configuration rulesProxychecks the proxy route temporarilyDirectchecks the direct connectionDiagnosticspreserves troubleshooting clues
Global Routing: Config, Proxy, and Direct
What it is: Global Routing determines how requests handled by Shadowrocket are routed. Config, Proxy, and Direct correspond to the Config, Proxy, and Direct options in the interface. They are not three servers or levels of connection quality. Changing the routing mode affects whether rules are used to make routing decisions; it does not automatically fix server credentials or your local network.
Where to find it: Open Global Routing from Home, check the selected mode, then return to Home and confirm the selected server. Config here is a routing mode; the Config tab at the bottom is for managing configuration files. They are related, but they are separate parts of the interface.
| Mode and interface label | How requests are handled | When to use it | What to keep in mind |
|---|---|---|---|
| Config | Handles requests according to the rules and policies in the active configuration. | For everyday routing based on domains, addresses, and other conditions. | Confirm the configuration is active, then check rule order and the final rule. |
| Proxy | Sends handled requests through the proxy route. | Temporarily check whether the proxy route works without the influence of routing rules. | Not every app or system request will necessarily pass through the client. Restore your usual mode after testing. |
| Direct | Sends handled requests over a direct connection. | Compare how the current network accesses a destination directly. | Direct does not turn off the connection. To disconnect, use the switch on Home. |
How to set it: If you already have a working rules configuration, start with Config. If a destination fails, note the current mode, retry the same destination in Proxy, then compare with Direct. Change only the mode, keeping the server and test destination the same. This helps distinguish a rule sending traffic down the wrong route from a route that is unavailable.
Rule-based routing: match first, then apply a policy
What they do: Rules combine match conditions with handling policies. DOMAIN matches a specific domain, DOMAIN-SUFFIX matches a domain suffix, and DOMAIN-KEYWORD checks for a keyword in the domain. IP-CIDR and IP-CIDR6 describe IPv4 and IPv6 address ranges; GEOIP classifies an address using a geolocation database. USER-AGENT depends on available request information, and not every connection provides the fields needed for a match. FINAL handles requests that did not match an earlier condition.
Where to find them: Open Config, confirm which file is active, then open that file to review its rules and policies. Rules are generally evaluated in order, so a broad condition near the top may match before a more specific one below it. Saving rules in another file does not mean the current connection is using that file.
DOMAIN-SUFFIX,example.com,PROXY
DOMAIN,blocked.example.com,REJECT
FINAL,DIRECT
This example illustrates syntax and order only: the first rule already matches blocked.example.com, so the more specific reject rule below may never run. To handle that subdomain first, move the more specific DOMAIN condition higher in the list. This is not a complete configuration to replace an existing file.
How to set them: Back up the original file in Config, then edit one rule related to your test destination. PROXY sends traffic to the proxy policy, DIRECT uses a direct connection, and REJECT blocks the request. If a rule uses a custom policy name, check that the policy exists and is currently available. Save the file, confirm it is still active, then test the same destination again.
What to keep in mind: Domain and address rules use different information, and DNS results can affect subsequent address matching. Don’t add REJECT rules in bulk just because one page looks wrong; images, sign-in pages, and API requests may use different domains. Check connection records to identify the destination that is actually failing.
Subscribe and Add Server: Manage sources separately from entries
What they do: An individual server entry stores details such as the address, port, protocol, and authentication. Subscribe stores an existing subscription source and can update its server entries in bulk. The fields differ across Shadowsocks, VMess, VLESS, Trojan, HTTP, SOCKS5, WireGuard, and Hysteria2. Don’t infer the protocol from a server name or enter one protocol’s parameters under another type.
Where to find them: View servers and subscription groups on Home, then use the add option to open Add Server. When adding a server manually, choose the type that matches your existing details. To add an existing subscription, use Subscribe and paste the complete URL provided by your service provider. You can import an existing QR code with Scan QR Code; review the imported entries afterward.
How to set them: Give each subscription a note that identifies its source. After the first update, check that the list contains entries, then test one server. Updating a subscription and selecting a server are separate actions: a successful update means the source was read, not that every entry can connect. For manual entries, check the address, port, authentication, and any TLS, SNI, or transport fields required by your existing details. Don’t guess values.
When you have multiple sources, group entries by source rather than relying on similar names to spot duplicates. Before removing anything, check whether an entry was added manually or generated by a subscription; local edits to subscription entries may be overwritten on the next update. Before deleting an old source, confirm the one you’re keeping updates successfully and its selected server works.
What to keep in mind: If the list is empty, first check that the URL is complete and that you haven’t pasted a webpage address instead. Then check the source status and your current network. Adding the same source again usually won’t fix a parsing problem and can make future maintenance harder. For the import steps, see the setup guide.
On Demand: Connect based on network conditions
What it does: On Demand lets the system start or manage a VPN connection according to configured conditions. It controls when to connect; Global Routing controls how requests are handled after connection. They are not interchangeable. Enabling On Demand does not select a working server or fill in missing rules.
Where to find it: Look for On Demand settings in Settings, and check which VPN configuration is currently in use by the system. Available conditions and options depend on what the device displays. Review network type and Wi-Fi conditions in the context of how you use the device.
How to set it: Connect manually first and confirm the server, Config, and DNS are working. Then enable On Demand with one clear condition, such as a connection behavior for a particular type of network. Switch to that network and check whether the connection behaves as expected; then check its state after leaving the network. Add other conditions only after the first one works.
What to keep in mind: If the connection comes back soon after you turn it off manually, an On Demand condition may have been met again; the switch may be working as intended. To troubleshoot automatic reconnections, temporarily disable On Demand, then observe the Home switch and system VPN status. On managed devices, some system behavior may be restricted by organizational policies; check the permissions available on the device.
Suggested test sequence
- Turn off automatic triggers and test one manual connection.
- Note the current network, selected server, and Global Routing mode.
- Enable one On Demand condition, switch networks, and observe the status.
- If the result is unexpected, undo that condition first; don’t change rules and DNS at the same time.
Data: Read traffic figures separately from connection results
What it shows: Data displays traffic measured within the client’s statistics scope. Readings can help show whether data was transferred during a connection and whether traffic was concentrated on an expected server or app. They are not a complete device usage report and do not replace usage records from your service provider. Different starting points, background connections, retries, and protocol overhead can lead to different readings.
Where to find it: Open Data and review the available statistics, such as server or app views. Check the time range first. The information available in the system and the interface may vary; a single label or reading cannot establish that every request from an app followed the same route.
How to use it: Note the current reading and time, return to Home to confirm the selected server, then visit a specific test destination. Check Data again and compare the change with connection records to assess the policy used. To compare Config and Proxy, use the same destination and similar actions both times. Avoid downloading a large file or running other sustained transfers during the test.
What to keep in mind: An increase only shows that data was transferred within the statistics scope; it doesn’t prove that every resource on a webpage loaded successfully. No change may reflect request reuse, caching, or the scope being observed. To check whether traffic followed the expected policy, review the policy shown for that connection. To check accessibility, make an actual request.
Before clearing statistics, note the time range and readings you need to keep. A screenshot showing only a cumulative total usually isn’t enough to troubleshoot an intermittent issue. Include the time, destination, network type, and routing mode so the change in statistics can be matched to what happened.
Settings: DNS, test methods, and useful tools
Settings options have different effects: some affect name resolution and testing, while others change shortcuts or data syncing. Note the original value before making changes, and check whether the current Config includes related settings. If the configuration file manages the same options, review the active configuration instead of relying on a single toggle in one screen.
DNS: Check name resolution before changing its source
What it is and where to find it: DNS resolves domain names to addresses. Review the related settings in Settings and the DNS configuration in the current Config. Failed resolution or results that don’t suit the network can make it seem like “the server test works, but some domains won’t open.” This is different from a proxy server authentication failure.
How to set it and what to keep in mind: Keep the current working configuration. Check whether the destination domain resolves and whether the result is expected, then choose a resolution method that fits your network and configuration. Change one source or related option at a time, then retry the destination. Changing DNS is not a universal fix; if the connection record shows an authentication failure, check the server settings first. Check IPv6 behavior against your current network rather than turning off multiple options as soon as an issue appears.
Test Method and Connectivity Test: Keep test conditions consistent
What they are and where to find them: Test Method determines how connection tests are run; check the current option in Settings. Connectivity Test, on Home, runs the test. Different methods check different stages of the connection, so readings from different methods should not be compared directly.
How to set it and what to keep in mind: Before comparing servers, use the same Test Method, test destination, and network conditions. If a test fails, retry it, then try accessing the destination you actually need. A lower test reading does not mean sustained transfers will be faster or that every destination is accessible. Verify real requests even after a test passes. If you set a test URL, confirm that the URL itself works.
Today Widget: Verify status after using shortcuts
What it is and where to find it: Today Widget is a system-provided shortcut. Check its options in Settings and add it in a widget area supported by the system. Its appearance and available actions depend on the system version and the widgets the client provides.
How to set it and what to keep in mind: First establish a working connection in the app, then add the widget and test its shortcuts. After using a widget action, return to Home and verify the server and connection status. A shortcut won’t choose the right subscription, routing mode, or configuration file for you. If the connection behaves unexpectedly, check it in the app.
iCloud Sync: Keep a copy before resolving conflicts
What it is and where to find it: iCloud features sync or access data as permitted by the system. Check system iCloud permissions first, then review Settings and the relevant import options. Import from Cloud JSON is one way to access cloud data; a single import does not mean all settings stay in sync.
How to set it and what to keep in mind: Check that both devices use the expected account and permissions, then confirm the sync scope in the interface. Keep a copy of the current working configuration before importing. After import, verify the source, entry count, and active file. If there’s a conflict, decide which data to keep before making further changes on multiple devices. Data syncing does not transfer system VPN authorization or keep connection states identical across devices.
Diagnostics: Preserve useful clues in chronological order
What it does: Diagnostics provides tools and records for troubleshooting. Useful diagnostics should answer three questions: Did the request enter the handling scope? Which policy matched? At what stage did it fail? “It doesn’t work” alone isn’t enough to distinguish local network, DNS, rules, server authentication, or destination response issues.
Where to find it: Check Diagnostics and available logging options in Settings, then compare them with connection records, Home status, and Data. Use the options shown in the current interface; don’t assume or report information that isn’t displayed.
How to use it: Note the network type, destination domain, Global Routing mode, and selected server when the issue occurs. Enable diagnostic logging if needed, then reproduce one specific action and note the time. Check the results from around that time. If you temporarily enabled more detailed logging, restore your usual settings after collecting what you need to avoid accumulating unnecessary records.
- Check the network first: Confirm the device’s current network works. If needed, compare using Direct.
- Then check the server: Verify the selected entry and authentication details, then run Connectivity Test using the same method.
- Next, check the rules: Compare Config with Proxy and see whether the expected policy matched.
- Finally, check DNS and the destination: Use the DNS results and the stage where the request failed to decide what to change next.
What to keep in mind: Diagnostic materials may contain subscription URLs, server addresses, authentication details, and browsing records. Before sharing, remove credentials and unrelated personal information while keeping the time, error type, and relevant configuration context. Don’t publish complete configurations or cloud export files. For further troubleshooting, see the FAQs or the complete manual.