01 / GET & INSTALL
Get and install Shadowrocket
Choose an installation channel for your device, then prepare the information needed to connect. The iPhone and iPad app is a one-time $2.99 purchase on the App Store; the Mac, Windows, Android, and Linux versions are free. Desktop and Android installers are available from this site at /dl/shadowrocket/. Choose from the all-platform downloads; don’t identify a file by a similar name in search results. Installing the app, importing server details, and establishing a connection are three separate steps. Installing the app does not provide a working proxy connection by itself.
iPhone / iPad: verify the App Store listing
On the App Store product page, check the name Shadowrocket, developer Shadow Launch Technology Limited, app ID 932747118, and official icon. Refer to the App Store page for system requirements. The displayed currency and availability depend on your account’s storefront. If you can’t find the app, check your account region, content restrictions, and payment status; don’t mistake a similarly named app for the same product. Change your account region through your device’s account settings, and read the notices there about your balance, subscriptions, and payment methods first.
If you’ve purchased the app before, make sure you’re signed in with the Apple ID used for that purchase, then look for it in your purchased items. Changing devices or reinstalling doesn’t mean you need to buy it again, but whether the listing appears depends on your account history and current store availability. You can redownload it on iPhone or iPad using the same purchasing account, subject to App Store rules. For steps to find your purchase, see the guide to restoring apps from your purchased list. After installation, don’t import all your settings at once; first confirm that the app opens normally.
Desktop and Android: choose the right installer
Mac has separate installers for Apple silicon and Intel. Check your chip or processor in About This Mac and choose the matching file; don’t guess based on the device’s appearance or purchase year. On Windows, use the matching installer. For Android, make sure the device architecture matches the download page. On Linux, choose deb, rpm, or AppImage according to your distribution’s package management. These formats serve different purposes and aren’t interchangeable by changing a file extension. Work or school device policies may also restrict installation, so check whether you’re allowed to install apps yourself.
| System prompt | Where to go | Before you proceed |
|---|---|---|
| macOS can’t verify the developer | System Settings → Privacy & Security → Open Anyway | Confirm the download came from this site |
| Windows SmartScreen | More Info → Run Anyway | Confirm you selected the intended installer |
| Android blocks the installation | Allow installation of unknown apps | Grant permission only to the source you’re using for this installation |
| Linux AppImage won’t run | Allow the file to run as an application | Confirm the file architecture matches your system |
These desktop and Android installers aren’t code-signed, so the first launch may trigger a system warning. Confirm the download source, then follow the steps in the table; don’t disable system-wide protections just to open one app. If Mac says the app is “damaged,” confirm the file has downloaded completely and is in Applications. Then use the command below to remove the quarantine attribute from that app and try opening it again. The command applies only to the specified app path, not the whole disk. Before running it, check that the path and app name match the actual installation location.
xattr -cr /Applications/Shadowrocket.app
For an AppImage on Linux, grant execute permission in the file’s properties or run chmod +x on the downloaded file. Use your system package manager to handle dependencies for deb and rpm packages. If installation fails, note the system message, processor architecture, and package format, then see the installation and security prompt guide. Don’t try files for different architectures at random. After installation, keep a note of the download source and package format so you can use the same channel for future updates.
02 / FIRST LAUNCH
First launch and permissions
On first launch, learn where things are before deciding which permissions to grant. The examples below use iPhone and iPad: Home shows servers and connection status, Config manages rule configurations, Data displays connections and traffic, and Settings controls client behavior. Items may move between platforms or interface updates, but server parameters, routing modes, and system connection permissions still serve different purposes. If you can’t find an option, look under its function rather than deleting the app to reset the interface.
Connection permission and server details are separate
When you connect for the first time, the system may ask to allow a VPN configuration. Confirm only when Shadowrocket has just initiated the connection and the app name in the system dialog is what you expect. Complete any device authentication the system requests. This permission lets the app manage network traffic at the system level; it doesn’t supply a server address, password, or subscription. If you tapped Cancel, you can usually return to the app and start the connection again, then follow the prompt. Don’t mistake a denied permission for a failed server, and don’t repeatedly change protocol fields.
If the system already has other network configurations, check which one is currently establishing the connection. Multiple network-routing configurations, device management policies, or system-level filters can affect one another. For a reproducible test, keep only the connection method you need, noting existing settings before changing anything. Managed devices may not allow VPN configurations to be added or changed; contact the device administrator, since switching servers in the app usually can’t override system policy. Troubleshoot permissions separately from network connectivity.
Grant additional permissions only when needed
Camera access is needed only when you use Scan QR Code. To scan a code from Photos, choose the minimum photo access required in the system prompt. QR codes can contain connection details such as passwords, keys, or subscription credentials; don’t share the scan screen in a public forum. Denying camera access doesn’t prevent you from entering details manually with Add Server, so you can verify the fields that way first. If you previously denied access, change it in the app’s system permissions page, then return to the original task instead of repeatedly trying to scan.
Local network permission depends on what you’re trying to access. If you need to connect to a server on your local network or reach a device at home, check whether the system allows that access. If you’re troubleshooting a remote connection, don’t attribute every failure to this permission. If only local addresses are unreachable, first confirm that the devices are on the same network, the target is powered on, and its service port is listening. Then check whether a rule is applying the wrong policy to local addresses. Verify permissions, addresses, and rules separately so you don’t change too many variables at once.
Make a short checklist before configuring
Before importing anything, note whether you’re on Wi-Fi or cellular, whether the system already has a VPN configuration, whether the app’s connection switch is off, and whether you plan to add a server manually or use an existing subscription. Don’t record passwords; note only the source of the information and the order of your actions. This helps identify whether an error occurred while granting permission, saving data, or connecting. If the app already has entries, check what they’re for. You can temporarily disable an unfamiliar old configuration and delete it later once you’ve confirmed it’s no longer needed, so useful rules aren’t lost.
Keep the default settings for now. Don’t rush to enable On Demand, add multiple Configs, or change DNS. First connect using a server with known, complete parameters, then add features one at a time. Note the original value before each change, and run the same verification once afterward. If something goes wrong, revert the last change rather than piling on more settings. This keeps the cause easier to isolate and prevents permission, configuration, and remote service issues from being lumped together as “not working.”
03 / SERVER & SUBSCRIBE
Add existing servers and subscriptions
First determine whether you have individual server details, a single server share, or a subscription address containing multiple servers. Individual details usually specify one protocol and a set of authentication fields; a subscription is an entry point for retrieving and refreshing a group of items. Both may look like links, but they shouldn’t be entered into the same field indiscriminately. Use only information from your own existing service. Before importing, make sure the content hasn’t been wrapped across lines, had spaces added, or been truncated by a messaging app. Check that the link is complete, especially at the end, and don’t publicly share its credentials after copying it.
Add Server: enter fields for the selected protocol
In Add Server, select the type first, then fill in its fields. Enter a hostname or IP address in the server field and the numeric port provided by the service in the port field. Don’t include a protocol prefix, web path, or port in the address field. A remark is just for identification; a correct remark doesn’t mean the connection parameters are correct. Fields that look similar aren’t interchangeable across protocols. Review each entry before saving it as a separate item. Don’t add several similar configurations at once, or you may test the wrong one.
| Protocol | Fields to check | Common pitfalls |
|---|---|---|
| Shadowsocks | Address, port, encryption method, password | The encryption method must match the server |
| VMess / VLESS | Address, port, UUID, transport parameters | Don’t guess transport or TLS parameters |
| Trojan | Address, port, password, TLS parameters | Authentication passwords and certificate names serve different purposes |
| WireGuard | Keys, endpoint, address, allowed IP ranges | Distinguish the local private key from the peer’s public key |
| Hysteria2 | Address, port, authentication, TLS parameters | The network must allow the required UDP traffic |
For TLS connections, the connection address and certificate verification name may differ. Use the values provided in your existing configuration. With WebSocket transport, the path and Host are also part of the protocol settings; a single missing character can cause the handshake to fail. Don’t disable certificate verification just to get a test result, and don’t fill every blank field with the server address. Check the original configuration for missing required parameters; the client can’t infer all server settings from the protocol name alone.
Subscribe: keep subscriptions and their entries grouped
To add an existing subscription, select the relevant subscription type, enter the complete address in the URL field, save it, and then update it. Wait for the list to load. Give the subscription a recognizable remark so you can identify its source and manage it later. After updating, check that server entries appeared, then select one to test. If the list is empty, check whether the address is still valid, the network can retrieve a response, and the content is in a format the client can recognize. A webpage opening successfully doesn’t mean its response is an importable subscription.
Updating a subscription usually regenerates the entries it manages, so changes made directly to a server inside the subscription may be overwritten next time it updates. For a temporary comparison test, keep the original subscription and edit a clearly separate entry; remove the temporary item when you’re done. If multiple subscriptions have entries with the same name, don’t delete them by name alone. Check the group, address, and protocol. For a safe cleanup order, see the guide to managing multiple subscriptions so you don’t delete a group that’s still in use.
QR codes, import results, and credential safety
Scan QR Code is useful for importing individual server details, but check the detected type and fields rather than relying on an “Import successful” message. A QR code can also contain a subscription address, so confirm whether the import created a server entry or a subscription group. Import from Cloud JSON and similar options are for their designated formats; don’t rename arbitrary text files and try to import them. Whatever method you use, reopen the saved entry and check its key fields to make sure an incompatible format didn’t omit authentication or transport parameters.
Treat a subscription address as an access credential. Screenshots, logs, and exported configurations may contain sensitive information. When describing an error to someone else, include only the protocol, error type, and field structure with sensitive values redacted. If an address was exposed accidentally, change the credentials through the original service and update the client entry. Importing only confirms that the data was saved. Next, select the right server and Global Routing mode, then test the connection to confirm the details work on your current network.
04 / GLOBAL ROUTING
Global Routing: three modes and rule order
Global Routing determines which policy applies to connections handled by the client. Config, Proxy, and Direct select configuration-based routing, the proxy path, and the direct path, respectively. They aren’t three servers or three protocols: a server entry describes how to connect remotely, while a mode determines when and where traffic goes. When troubleshooting, note both the current mode and selected server. Otherwise, different results for the same website across modes can look like random connection changes. Understand what each mode does before adding complex rules; it’s usually easier to get consistent results that way.
| Mode in the interface | How it works | What to test |
|---|---|---|
| Config · Configuration | Applies policies based on rules in the active configuration | Check the matching rule and final policy for the destination |
| Proxy | Sends handled connections through the proxy route | Check whether the selected server can reach the destination |
| Direct | Sends handled connections through the direct route | Establish a baseline for local network access |
Selecting Config doesn’t automatically enable a file
After selecting Config, check which configuration is actually active on the Config page. Importing a file, editing it, and enabling it are separate actions. A configuration may contain rules, policy groups, DNS settings, and other content; its policy names must match available policies. If a file refers to custom policy groups that aren’t defined locally, it may not behave as expected even if the syntax looks complete. Before using a configuration for the first time, review its main sections. Don’t replace your existing settings wholesale without understanding its default policies.
From specific rules to the fallback
Rules are usually matched in order, so put specific conditions before broader ones and use FINAL as the last-resort rule. DOMAIN matches a full domain, DOMAIN-SUFFIX matches a domain suffix, and DOMAIN-KEYWORD matches a keyword in the domain. Keyword rules can match too broadly; don’t add common short words just because one request behaved unexpectedly. IP-CIDR and IP-CIDR6 match address ranges, while GEOIP relies on IP geolocation data. Domain and address rules use different matching inputs and aren’t interchangeable.
[Rule]
DOMAIN-SUFFIX,example.com,PROXY
IP-CIDR,192.168.0.0/16,DIRECT
GEOIP,CN,DIRECT
FINAL,PROXY
This example only illustrates rule order; it isn’t a complete configuration for every network. The first line sends the example domain to PROXY. The second routes the listed local network ranges through DIRECT. Next, traffic is routed by IP geolocation, and FINAL handles any connections that didn’t match earlier rules. PROXY and DIRECT are rule policy keywords; keep them distinct from the interface labels Proxy and Direct, including their capitalization. When copying, avoid full-width commas or extra quotation marks. After saving, confirm the file loads before testing the destination.
Local networks include more than the example range; build rules for your actual network, and consider IPv6 separately. If a request starts with a domain, DNS results may affect how address rules are applied. So if an IP is routed directly, don’t infer the matching process from the domain suffix alone. Check the destination, matching rule, and policy in the actual connection record to see whether a domain rule matched or the request reached address matching or FINAL. If there’s no record, start with a small set of explicit rules to establish a baseline.
Isolate issues with the three modes
If a destination works in Proxy but fails in Config, check rule order, policy group selection, and DNS settings in the active configuration before rebuilding the server entry. If it also fails in Proxy, check the server parameters, current network, and remote service. Direct provides a comparison for the direct route; it doesn’t guarantee every destination will be reachable. Change one mode at a time while keeping the server, network, and destination the same so results are comparable. Restore your usual mode after testing.
Before changing rules, make a copy of the configuration as a rollback point. Add only the rule needed to address the current issue, and note why you added it. More rules don’t necessarily mean better matches; duplicates and overlapping broad conditions make maintenance harder. Conditions such as USER-AGENT also have limits—don’t assume every encrypted connection exposes application-layer information that can be matched. Don’t enable extra decryption settings just to make a condition match. First check whether existing domain, address, and policy rules already solve the problem.
05 / CONNECT & VERIFY
Connect and verify step by step
Start with the shortest path: confirm the local network works, select one server, note the current mode, and turn on the connection. If a system permission prompt appears for the first time, follow Chapter 2. Check whether the app stays connected and whether an error appears immediately, but don’t treat an enabled switch as proof that the connection works. The full path includes the server handshake, DNS resolution, destination connection, and the app’s request. A failure at any layer can leave a webpage waiting, so diagnose each layer instead of repeatedly toggling the connection.
Understand what a test result tells you
Interpret Connectivity Test results in the context of the selected test method. Different tests may cover only network round trips, port connectivity, or a particular HTTP request; they may not follow the same complete path as your actual use. A latency result doesn’t prove every authentication field is correct, nor does it guarantee sustained throughput, video playback, or a specific app will work. No result may also be due to the test target or method. Record the test method and network, then compare results under the same conditions. Don’t rank numbers from different kinds of tests against one another.
For the first verification, use one server with clear, complete settings. In Proxy mode, open a familiar webpage and confirm it shows fresh content rather than a cached page. Then test the app you actually need, and finally switch to Config to check its routing. After changing modes or servers, make a new request. Some apps reuse existing connections, so an old connection may keep using its previous route after settings change. If needed, close and reopen the target app; you don’t have to restart the whole device for every test.
Troubleshoot by symptom
If all destinations fail after connecting, disconnect and check whether the basic network works. Then check the selected server, authentication parameters, and system VPN status. Changing rules won’t fix an unavailable underlying network. If only some domains fail, check for incorrect domain rules, unexpected DNS responses, or an issue with the destination itself. If webpages work but a particular app doesn’t, check the app’s domains, protocols, and existing long-lived connections. Don’t treat a successful browser test as proof that the app works.
Handle timeouts and authentication failures separately. For a timeout, first check the address, port, network reachability, and whether the remote service is listening. For an authentication error, recheck the password, UUID, key, or protocol fields. For TLS errors, check the device time, certificate name, and transport settings; don’t respond by disabling verification across the board. If a subscription was just updated, confirm that the selected entry hasn’t changed. Otherwise, you may keep editing an old entry while the connection is using another one.
| Reproducible symptom | Check first | Next comparison |
|---|---|---|
| No internet even when disconnected | Basic network, captive portal | Restore basic connectivity, then retry |
| Proxy works, Config fails | Rules, policies, DNS | Check the matching result for the destination |
| Wi-Fi fails, cellular works | Network restrictions, DNS, UDP reachability | Retest with the same server |
| Webpages work, one app fails | App requests and long-lived connections | Reconnect the app and check its connection records |
Keep test conditions the same when switching networks
Switching between Wi-Fi and cellular can help determine whether an issue depends on the network. Keep the server, mode, and destination the same, and wait for the system’s network status to stabilize before testing. Public Wi-Fi may require a captive portal sign-in. Complete that first, then connect the client. Don’t repeatedly update subscriptions or change DNS before signing in; that can make a basic network issue look like several features have failed at once and add misleading error records.
Verification should also check stability over time. A webpage loading once only shows that a route worked at that moment. Use the client normally for a while and check for repeated reconnects, disconnects after the screen locks, or issues limited to large downloads or UDP apps. Different symptoms point to different layers: for lock-screen issues, check system behavior and On Demand; for specific transport issues, check the protocol and network support. Don’t judge by a single latency result. For more on test limitations, see latency tests and real-world access.
06 / DATA & CONNECTIONS
Data: traffic and connection records
Data helps you understand what happens during a connection; it doesn’t replace records from a remote service account. Before reviewing traffic, identify whether the page shows a current rate, cumulative traffic, or data for an individual connection, and when the statistics period started. Different screens and metrics may cover different scopes; a displayed value shouldn’t be treated as the device’s total carrier-billed usage. Instead, compare a fixed time window and consistent actions to see whether a request was recorded, which policy it used, and whether data continued to transfer.
Separate transfer rate, total usage, and connection count
A transfer rate shows changes over a short period, cumulative usage totals traffic over a reporting period, and connection count describes requests or sessions. They don’t map directly to one another. A long download may use only a few connections while transferring a lot of data; a webpage can create many connections to different domains for images and background requests. A high connection count alone doesn’t indicate a problem. Consider which apps and destinations were active, and check whether the activity fits your use before investigating further.
Upload and download are usually labeled from the device’s perspective, but check the specific metric and its scope. Background photo sync, system updates, cloud storage, and media preloading can transfer data even when you’re not actively using an app. If usage looks unusual, pause known high-traffic tasks, close unrelated apps, and observe the change from a single destination. Don’t assume a brief spike means traffic is leaking. A counter resetting after you refresh the page also doesn’t mean earlier traffic never crossed the network.
Use connection records to verify rule results
When troubleshooting rules, connection records explain more than whether a webpage opened. Where available, check the destination host or address, matching rule, final policy, selected server, and error. Make a new, easy-to-identify request, then find its entry in the records. If an app connects to several domains at once, check what each one is used for. Don’t infer that the whole app is connecting directly just because one entry shows DIRECT. Login, content loading, and media playback may use different destinations.
A full domain name may not appear in every record. Some requests use an IP directly, or cached data may affect the time between DNS resolution and connection, so records may show only an address. Use the request time, target app, and rule conditions to interpret it; don’t add a broad blocking rule based on one unfamiliar address. If a policy still looks unchanged after editing the configuration, confirm which file is active, whether it was saved, and whether you created a new connection. A reused connection won’t be evaluated against the latest rules.
Understand differences between client and other traffic statistics
Differences between client totals, system network usage, and remote account traffic don’t necessarily mean there’s an error. The time range, whether direct traffic is included, protocol overhead, retries, and the remote service’s measurement method can all affect totals. To compare figures, align the reporting periods and make sure you’re comparing the same scope. Don’t use a client total to calculate the remaining allowance on a remote account; check the records in your own account. Data is for observing behavior and troubleshooting, not billing.
To verify a rule, note the current state, perform just one action—such as reopening a webpage—and check the new connection record. Then resume normal app use. A short, controlled test is easier to interpret than searching through records while many background tasks are running. If a request appears to be retrying in a loop, stop the target app or disconnect to prevent a flood of repeated requests, then investigate the error. Don’t let a failing request keep retrying while repeatedly clearing the records; that makes useful context harder to find.
Remove sensitive information before sharing records
Connection records may reveal destinations, internal network addresses, server details, and activity patterns. Detailed logs may also contain credentials from your configuration. Before sharing a screenshot, crop unrelated areas and redact sensitive information, leaving only the necessary error and fields. After troubleshooting, turn off extra detailed logging if you no longer need it. If you reset cumulative statistics, note the reporting period and reset time first. Resetting changes only the local record; it doesn’t undo network traffic already sent or change remote account usage.
07 / SETTINGS
Settings, DNS, and On Demand
Settings control client behavior, but they can’t replace correct server parameters and rules. Before changing anything, identify your goal: automatically connecting on certain networks, resolving domains differently, or changing how a test runs. If you can’t identify the symptom you’re trying to fix, leave the current value as is. Options may be located differently across platforms, and some settings may be controlled by a configuration file. Before editing, check which source controls the current value so an active configuration doesn’t overwrite a change made in Settings.
DNS: check the resolution result before changing how it works
DNS translates domain names into addresses; a proxy determines the subsequent connection route. They’re related, but not the same. If a domain can’t be resolved, the server connection may be fine while access still fails. Even with a resolved address, the connection can fail because of a rule or remote service issue. First determine whether DNS returned no result, an unexpected result, or a result followed by a connection timeout. Don’t blame every slow load on DNS, or assume a resolver that works on another network will work with your current network and rules.
When choosing a DNS method, consider the route used for DNS requests, whether the resolver is reachable, and how the result may affect subsequent address rules. Using a resolver that requires the proxy to reach it for the DNS lookup needed to establish that proxy connection can create a dependency loop. First confirm that the server address can be resolved on the current network, then verify from simple, explicit settings. Change one DNS option at a time, make a new request, and check the records. Allow for cached results to change before drawing conclusions.
When mixing domain and IP rules, DNS responses, IPv4 and IPv6 addresses, and cache state can all change which rule matches. If only one address family fails, check whether the current network can reach it and review the address rules in your configuration; don’t apply the same fix to every request. After changing settings, test common websites, local network devices, and the target app to make sure you haven’t fixed one destination by changing the route for others. Keeping a recoverable copy of the original configuration is better than making undocumented changes one after another.
On Demand: verify manual connections first
On Demand triggers connections when configured conditions are met; it doesn’t fix incorrect server details. First confirm that a manual connection is stable, then configure On Demand one condition at a time. Check the scope and order of the conditions—for example, whether the network type or named network is what you expect. Enabling the main switch alone doesn’t mean it will connect automatically in every situation. After saving, test both networks where it should trigger and networks where it shouldn’t. Check the system VPN status and actual requests to confirm the conditions work as intended.
When On Demand is enabled, manually disconnecting may not last: the system can reconnect when its conditions are met again. That doesn’t necessarily mean the switch is broken. First check whether On Demand is still enabled and review its matching conditions. To troubleshoot or restore basic connectivity temporarily, disable On Demand before disconnecting manually; re-enable it afterward according to your notes. Network changes, screen locking, and closing an app are different events, so testing only one doesn’t establish that all scenarios work. If a management policy is involved, distinguish system restrictions from app conditions.
Test methods, logs, and other options
Connectivity Test settings affect the test target and what its results mean. After changing the test method, establish a new baseline; don’t draw conclusions by comparing results measured differently. Enable detailed logs only when needed. Note the order of events, reproduce the issue once, and check the corresponding error. Then return logging to its normal level. Options involving certificates, decryption, or request rewriting aren’t required for a first connection. Don’t enable them just to make rules seem more complete without a clear need and permission, and never install a certificate from an unknown source.
After changing settings, return to Home and confirm the selected server and mode are unchanged. Test a fixed set of destinations, then check the actual policy in Data. If a change didn’t improve the issue, revert it before trying the next option; don’t leave an experiment in place indefinitely. Keep a brief change log with the original value, reason, and test result—never credentials. For help with a specific issue, use the FAQ to find matching symptoms instead of changing every setting that seems relevant.
08 / MAINTENANCE
Subscription updates, backups, and maintenance
The goal of maintenance is to keep a configuration that you can understand and restore, not to keep adding servers and rules. Once the connection is stable, note the active subscription groups, enabled configuration, usual mode, and any essential settings that differ from the defaults. Record their purpose and relationships, but don’t put authentication details in ordinary notes. Before making broad changes, confirm the current connection works and create a rollback point. If an update causes problems, you’ll then know whether they were already there or were introduced by the change.
Update subscriptions and the app separately
A subscription update retrieves the server list and related details from your existing subscription. An app update changes the client itself; they’re separate operations. If a subscription update fails, keep the existing entries and don’t immediately delete and re-add the whole group. Check whether the current network can reach the subscription address, whether the address is still valid, and whether the response is usable. If an existing server still works, keep using it while you troubleshoot the update. Deleting a subscription may also remove the entries it manages, leaving you without a working comparison.
After an update, check the number of groups, whether your commonly used entries are still there, and whether the selected entry changed. Entries with the same name may have different addresses or protocol parameters, so verify one with an actual connection test. Clearly mark temporary test entries you added manually and remove them when you’re done to avoid confusing them with subscription entries. Organize multiple subscriptions by source; don’t use a region name as the only identifier. Before deleting duplicates, check their group and purpose rather than removing every entry with a similar name.
Keep your data recoverable when updating the client
Manage iPhone and iPad app updates through the App Store. For Mac, Windows, Android, and Linux, get the appropriate updated installer from this site’s download page. Before updating, save any configuration you need and confirm it can be restored. Don’t delete the app and all its data just to update it. On desktop, quit the running client first and use the correct architecture and package format. If the new version triggers a system warning on first launch, follow the source checks and steps in Chapter 1; don’t skip source verification just because you’ve installed the app before.
Backups should cover server details, existing subscription addresses, the active Config, and important settings. Check the export screen to see exactly what’s included; don’t assume a single file contains everything. After exporting, confirm the file can be read and save it somewhere secure that you control. Configuration files may contain passwords, private keys, and subscription credentials, so don’t put them in public cloud links or forums. When moving to another device, restore only the essential data first and verify the connection. Migrate the remaining rules afterward to make compatibility issues easier to isolate.
Moving devices, restoring data, and handling long-term failures
After changing devices, get the app first, then restore your data. The iPhone and iPad app is a one-time $2.99 purchase on the App Store; Mac, Windows, Android, and Linux versions are free. If you’ve purchased the iPhone or iPad app before, sign in with the original purchasing account and find it in your purchased items. For other platforms, download it again from this site. Purchase history can restore access to the app, but it won’t automatically restore all local settings or your remote service details. Manage the app, backups, and remote account separately; they aren’t the same set of data.
For a persistent failure, retrace the same steps in order: check the basic network and system time, then connection permissions, the selected server, Global Routing, the active configuration, and DNS. Finally, verify whether the remote parameters have changed. Keep each symptom reproducible, and if you recently updated or changed rules, compare against the last known working state. Don’t reinstall the app, erase its data, and replace all parameters at once. Even if the connection comes back, you won’t know what fixed it, and you may lose information you still need.
Routine checks don’t require frequent resets, repeated speed tests, or constant server switching. Run a targeted test only when there’s a new symptom, the network has changed, or you’ve updated a configuration. If you need help, gather the operating system, network, selected mode, error type, and steps already tested, with sensitive information removed. To retrace the main steps, return to the quick start guide. For App Store search issues, see the App Store search troubleshooting checklist; for other common issues, browse the FAQ by topic.