Shadowrocket Official Manual · Configuration Reference

Shadowrocket Rules and Configuration Guide

This page is organized by configuration section, rule field, and troubleshooting workflow. Use it when editing or reviewing a config. If you’re importing an existing subscription, choosing a connection mode, or turning on the connection for the first time, start with the Getting Started guide. If you’re connected but want to know why a request went direct, through a proxy, or was rejected, jump to the relevant section below.

Examples use reserved domains and documentation addresses to explain configuration and matching behavior; they are not working routes. Before testing a proxy policy, make sure your existing subscription or server details are valid. This site does not provide, sell, or recommend nodes or subscriptions.

Configuration Structure: Settings, Rules, and Mappings

Check which file is active first

A configuration file can contain multiple sections marked by square-bracketed headers. Each header determines how the following lines are interpreted, up to the next section header. The same text can mean different things in different sections. For example, portal.example.com can be a hostname to match in [Rule], or a hostname to map in [Host]. Find the section before editing its contents. Don’t treat the file as one long list of interchangeable rules, and don’t insert a new section header in the middle of an existing rule list.

Check the selected configuration in Config. If the list contains files with the same name, it can be hard to tell which one you just edited. Keep a working original and give the test copy a name that makes its purpose clear. Editing, saving, and selecting are separate actions: saving a copy doesn’t mean traffic is using it. After making changes, confirm that the copy is selected and that Global Routing is set to Config. If the original file comes from a remote URL, refreshing it may overwrite local changes. Keep customized rules in a separate copy if you need them to persist.

SectionPurposeCheck when
[General]Controls overall network behavior, such as DNS settingsYou can reach an address, but domain resolution fails
[Rule]Selects a policy based on conditionsThe actual policy differs from what you expected
[Host]Maps a specific hostname to an addressOnly certain hostnames resolve to unexpected addresses
[URL Rewrite]Rewrites visible URLs that matchRedirects or path-based handling behave unexpectedly

[General] and [Host]: DNS resolution is not routing

[General]
dns-server = system

[Host]
portal.example.com = 192.0.2.10

Here, dns-server = system specifies the system DNS settings. Actual results also depend on the current network, cache, and how the app resolves names. The [Host] example maps a specified hostname to a documentation address. This is not a routing policy, and it won’t make a service at that address available. In a real environment, consider a static mapping only after confirming who owns the target address, what it is for, and who maintains it. If the target rotates across servers or relies on dynamic DNS, a static mapping may start failing after its address changes.

A static address also won’t change how the browser validates a certificate for the original hostname. Pointing a hostname to a server that accepts connections but has no certificate for that name can still cause a name mismatch. If a test report only says “can’t open,” separate the failure into DNS resolution, connection setup, and certificate validation. Success at one stage only means the request reached the next; it doesn’t pass that stage for you. After removing a test mapping, send a new request rather than judging the change from a page that was already open.

[Rule] and [URL Rewrite]: Don’t mix them

[Rule]
DOMAIN,portal.example.com,DIRECT
FINAL,PROXY

The first rule sends the exact hostname direct. The last rule sends other traffic handled by the rule list through a proxy. [URL Rewrite] uses a different syntax: don’t put a path-matching regular expression in the DOMAIN field, or use a redirect action in place of PROXY. For an unfamiliar file, find the section first, then check the keywords, separators, and policies. A line that looks like a domain alone doesn’t tell you whether it affects DNS, routing, or request handling.

Validate a configuration copy with the smallest possible change. Duplicate an existing config, leave DNS settings untouched, put the rule to test in the correct section before the fallback, and check that the file loads. If saving produces a parse error, start with the most recently added section header, full-width punctuation, or a line broken by the editor—not DNS or the server. A line wrapping on screen is only a display effect; a newline in the file can split one rule into two invalid records. To roll back, switch to the original instead of repeatedly undoing edits in the same file.

A subscription provides the information needed to connect; a configuration describes how requests are handled. After updating an existing subscription, still check that the intended config is selected and that the servers required by its policies are available. Neither replaces the other.

Global Routing and Policies: Choose a Mode, Then Read the Rules

How to Troubleshoot the Three Modes

Global Routing determines whether traffic handled by the client is routed according to the current configuration. The Config mode corresponds to Config and is used to test rules. The proxy mode corresponds to Proxy and provides a proxy-path comparison. The direct mode corresponds to Direct and provides a direct-path comparison. This applies to traffic handled by the client, not necessarily every kind of traffic on the device. After comparing the other two modes, switch back to Config to test individual rules.

Interface labelWhat it usesQuestion it answers
ConfigRules and policies in the current configurationWhich rule matched this request?
ProxyUses the current proxy pathDoes the proxy path work for this target without the current routing rules?
DirectUses a direct connectionCan the current network reach the same target directly?

For a comparison, keep the network, target, and server selection the same; change only the mode. If you also switch from Wi-Fi to cellular or choose a different server, it’s hard to tell whether the result changed because of the routing mode or the network. Don’t judge by a page that has already loaded: the browser may reuse an existing connection or display cached content. Start a new connection to the same target and check the visible request records. That’s more reliable than simply seeing whether the page loads. For repeated tests, record the mode and result each time rather than relying on memory.

PROXY, DIRECT, and REJECT

[Rule]
DOMAIN,admin.example.com,DIRECT
DOMAIN,metrics.example.com,REJECT
DOMAIN-SUFFIX,example.com,PROXY
FINAL,DIRECT

DIRECT connects to the target without a proxy server, but still depends on working DNS and a local network path. PROXY uses the proxy path, provided the existing server’s address, protocol, and authentication details are valid. REJECT rejects requests that match the rule. The app may show a connection failure, retry, or missing functionality rather than a clear rejection message. Configurations use these uppercase policy names; the interface’s Proxy and Direct options are mode settings. They operate at different levels.

For example, in the configuration above with Config selected, the admin hostname matches the first rule, the metrics hostname matches the second, and only other matching subdomains try the proxy. If switching to Proxy restores access to a target, that only shows the proxy path works for that test target. It doesn’t prove the original rule matched DIRECT; the request may have matched REJECT or failed at another stage. Likewise, if a rule correctly matches PROXY but the page still fails to load, that alone doesn’t mean the rule is ineffective.

Track Connection, Test, and Request Results Separately

Turning on the connection means the client’s connection state has changed; it doesn’t mean every target is reachable. Connectivity Test can help compare reachability through the current server, but its test target isn’t the same as an actual app request. A site may load its main document but fail on authentication, images, or API requests. Reproduce the failing action in the same mode, then look for failed requests at the same time. Treating the homepage as the only success criterion can miss problems with checkout, login, or uploads, which may open connections later.

On Demand controls the conditions under which a connection is established or maintained; it doesn’t assign policies to domains. If the connection restarts after switching networks, check the connection behavior first, then check the current mode and rules. This separates “Is traffic being handled by the client?” from “Which path does it take?” instead of trying to solve both with a single FINAL rule. If the device is managed by an organization, system policies may also restrict connection behavior; follow the device management requirements.

Domain Rules: Exact Hosts, Suffixes, and Keywords

Rule fields take hostnames

To route by domain, extract the hostname from the URL. For https://api.example.com:443/v1/items, the hostname is api.example.com; the scheme, port, path, and query parameters aren’t part of the domain match field. Use DOMAIN for a single, specific hostname; DOMAIN-SUFFIX to match a domain and its subdomains; and DOMAIN-KEYWORD to match a string appearing in the hostname. The latter two can catch requests you didn’t intend to include, so list examples that should not match before changing them.

[Rule]
DOMAIN,login.example.com,DIRECT
DOMAIN-SUFFIX,example.com,PROXY
DOMAIN-KEYWORD,preview-build,DIRECT
FINAL,PROXY
ConditionExample matchExample non-match
DOMAIN,login.example.comlogin.example.comtest.login.example.com
DOMAIN-SUFFIX,example.comexample.com、api.example.comnotexample.com、example.com.example.net
DOMAIN-KEYWORD,preview-buildpreview-build.example.netpreview.example.net

A suffix is matched at a domain boundary, not by simply checking the last few characters. You don’t need to add *. to DOMAIN-SUFFIX to include subdomains. A keyword rule, by contrast, doesn’t need to match a complete subdomain label and doesn’t guarantee that the hostname belongs to a particular company or project. If using these rules for a temporary cross-domain test, document the test and remove the rule afterward. For long-term rules, prefer verified hostnames or suffixes wherever possible.

Get the domain from the actual request, not the page name

Opening a page in the address bar doesn’t mean every request goes to the hostname shown there. Login flows may redirect, static files may come from another domain, and a page may load data from different APIs. If only one button fails, repeat the action to narrow it down, then check which host requests failed at that time. Add an exact rule for a confirmed host and repeat the same action. Use a suffix rule only when multiple subdomains really need the same handling. Sending an entire domain through another policy may make the issue disappear, but it won’t show which request needed the change.

If the request record shows only a target address, a domain rule may not have a hostname to match. When an app connects directly to an address, uses its own DNS resolution, or reuses an existing connection, don’t assume the client saw the domain just because the service normally uses it. Conversely, a DNS alias or multiple returned addresses don’t mean you should add both the alias and every address to the rules. Prefer request attributes that are actually visible to the client and relatively stable. For targets with frequently changing addresses, maintaining fixed IP rules is often less reliable than matching the hostname.

Include at least one non-match in every test

After setting a login host to connect directly, confirm it uses DIRECT and that another host in the same domain still follows the expected path. For suffix rules, test the root domain, a matching subdomain, and a similar host outside that suffix. These examples check match boundaries; they need not point to real services. For actual testing, use a target you can access and are authorized to test, and note the expected and actual policies.

If two rules use the same policy, the access result alone won’t tell you which one matched. Check existing match records, or use a temporary config copy with distinguishable policies for the condition under test and the fallback, on a test target you’re permitted to use. Restore the original settings after testing. Don’t apply a reject policy to an entire domain used in everyday business just to test one rule. If other interfaces fail too, it will be harder to tell whether the problem was pre-existing or caused by the test rule.

IP Rules: IP-CIDR, IP-CIDR6, and GEOIP

Prefix length sets the range, not the port

[Rule]
IP-CIDR,192.0.2.10/32,DIRECT,no-resolve
IP-CIDR,198.51.100.0/24,PROXY,no-resolve
IP-CIDR6,2001:db8:40::/48,DIRECT,no-resolve
GEOIP,CN,DIRECT
FINAL,PROXY

Use IP-CIDR for IPv4 and IP-CIDR6 for IPv6. The number after the slash specifies the prefix length in bits. For IPv4, /32 represents a single address, while 198.51.100.0/24 covers addresses from 198.51.100.0 through 198.51.100.255. Matching this range doesn’t mean every address in it is reachable as a host. A single IPv6 address uses /128; the example’s /48 covers addresses that share the first three hexadecimal groups. If you need an exception for one server, don’t widen it to an entire subnet for convenience.

When managing a dual-stack target, first confirm which address family the connection uses. The browser or system may choose either one, and an IPv4 rule won’t automatically cover IPv6 requests. If the same target sometimes matches an IP rule and sometimes falls through to the default, compare the destination addresses used by each connection, then check whether DNS results vary by network. Widening an address range just to make one test pass may also change unrelated requests to other hosts in that subnet. If the client can see a stable hostname, a domain rule often describes the intended behavior more accurately.

no-resolve controls resolution for this rule

When a request uses a domain name, an IP rule needs the destination address to compare it with a prefix. At the end of a rule that supports it, no-resolve means not to actively resolve the domain just for that rule; an already known address can still be checked. It isn’t a global DNS switch and doesn’t erase existing resolution results. The prefix rule above means “classify by address only when the destination address is already available.” Later rules may still need DNS, and establishing the connection may still require a DNS lookup.

To check this option, observe requests made directly to an address separately from requests made by domain name. Don’t conclude that an IP rule failed just because a browser can’t open an HTTPS address: the certificate may not identify that address, or the service may need the original hostname to return content. Record the matched policy separately from the service response. If removing no-resolve changes the result, check whether the domain finally resolved to a matching address or whether DNS queries became slower or started failing. Those need different fixes.

GEOIP classifies addresses; it doesn’t prove location

GEOIP,CN,DIRECT selects a direct connection when the current data classifies the destination address as CN. It uses address classification data, not the domain suffix, the client’s location, or the account’s region. The same domain may resolve to different addresses on different networks, and classification data may change with updates. Because this classification is broad, place the rule after explicit domain and subnet exceptions. Don’t use it to specify the path for one particular hostname.

For example, a specific service that should use a proxy may resolve to an address classified as direct. If the geographic rule comes before the service exception, the later, more specific rule never gets a chance to match. When troubleshooting, record the correct hostname, the address used for the connection, and the first matching rule. A domain suffix alone can’t tell you how the address was classified or explain why the same domain took a different path on another network.

Use address-classification rules only when you specifically need to route by address category. For an internal service, first verify the target subnet and the current network. DIRECT doesn’t create a route to an internal network.

Request Attributes and Rewrites: USER-AGENT, REJECT, and URL Rewrite

USER-AGENT matches a visible request attribute

[Rule]
USER-AGENT,ExampleClient*,DIRECT
FINAL,PROXY

USER-AGENT matches a request identifier the client can see. The example uses a wildcard pattern for identifiers starting with ExampleClient; it doesn’t look up a process by the app’s display name in the system. Web pages, API requests, and background connections from the same app may use different identifiers, while multiple apps may use the same request component. Before changing the config, verify that the identifier is present and stable in a test request you control. Then send a comparison request with a different identifier to confirm it isn’t matched too.

An ordinary encrypted connection doesn’t expose HTTP request headers just because its destination domain is visible. Don’t treat USER-AGENT as a way to identify apps across all HTTPS traffic, or assume that all requests from an app use the same policy. If the rule doesn’t match, first check whether the client can see the field, then verify the pattern and rule order. Widening the wildcard won’t make an invisible field available and may match unrelated requests. For a stable destination host, use a domain rule that’s easier to verify.

Use REJECT for hostname-level blocking

[Rule]
DOMAIN,telemetry.example.com,REJECT
DOMAIN-SUFFIX,example.com,DIRECT
FINAL,PROXY

This example rejects one specified hostname while allowing other hosts with the same suffix to proceed to the next rule. It doesn’t inspect paths or distinguish between interfaces on the same host. If a service uses one host for both login and metrics reporting, rejecting by domain may affect both. To assess the impact, start with an exact hostname in a config copy, then check the main page, login, and related interfaces. Don’t begin by rejecting the entire suffix. For retry behavior, false positives, and failure symptoms, see How REJECT Rules Work and Where They Apply.

Rules block requests that enter the matching process; they can’t undo a completed download or delete content already saved by an app. If the first test seems unchanged, the app may be using cached content. If repeated tests produce more connection alerts, the app may be retrying automatically. Interpret both cases alongside request timestamps and match records. If an error disappears but the feature also stops working, don’t count that as success. Repeat the same acceptance test you used before the change.

URL Rewrite works with visible URLs

[URL Rewrite]
^http://example\.com/old$ http://example.com/new 302

This example consists of a match expression, a replacement URL, and a redirect action, from left to right. The expression uses start and end anchors, and the escaped periods match literal dots, so it applies only to the exact example URL. A URL with query parameters won’t match this pattern. 302 illustrates a temporary redirect; the destination must be accessible on its own. Don’t include the destination URL in the same rule’s match scope, or testing could create a redirect loop. Avoid using a permanent redirect for trial and error: the browser may cache the result, making retests after rolling back hard to interpret.

Plain HTTP is used here to show how the URL field is matched; modern browsers may upgrade it to HTTPS, which won’t match the example. For HTTPS, path matching is possible only when the current handling environment can actually see the URL field. Seeing the hostname doesn’t mean the path or query parameters are visible. This guide doesn’t require you to configure decryption to test routing. If you only need to choose a connection path for an entire host, use [Rule]. If an app won’t follow the rewritten URL, first check whether the destination is unreachable, the request method changed, or the app doesn’t accept redirects. Don’t hide the problem with a broader expression.

Rule Order: Put Exceptions Before the Rules They Override

Check the first match; don’t guess priority from rule names

With Config selected, read the ordinary [Rule] list from top to bottom; the policy is determined by the first rule whose conditions match. If a suffix rule already catches a request, a more specific hostname rule below it won’t automatically take precedence. A later reject rule doesn’t gain priority just because its policy is REJECT. Before making changes, ask two questions: Which requests do you want to affect? And could an earlier condition already catch them?

[Rule]
DOMAIN,auth.example.com,DIRECT
DOMAIN,metrics.example.com,REJECT
DOMAIN-SUFFIX,example.com,PROXY
GEOIP,CN,DIRECT
FINAL,PROXY

In this example, requests to the known hostname auth.example.com go direct first, metrics.example.com are rejected, and other matching hosts use the proxy. If the suffix rule is moved to the top, both of the first two hosts match it first. The geographic rule comes after the suffix rule so that changes in address classification don’t affect these confirmed domains. The point isn’t that “domain rules always outrank IP rules.” It’s to handle clearly defined application-level exceptions first, then broader requests with conditions that are more likely to change.

FINAL only applies when nothing above it matches

FINAL takes no domain or address argument; put the policy immediately after the comma. For routine maintenance, keep one explicit fallback at the end of the list. FINAL,DIRECT sends requests without a matching rule direct by default, which suits configs that proxy only explicitly listed targets. FINAL,PROXY sends unmatched requests through a proxy, which suits configs with explicit direct exceptions. Changing the fallback changes the default behavior of the entire configuration; it’s not a local fix for one failing host. Before changing it, identify any unlisted services that still need to work.

For example, with FINAL,PROXY, a target may use the proxy even if a specific proxy rule never matched. Assuming that “it uses the proxy” means “the new rule worked” is a common misdiagnosis. Check the actual match, or use a test copy to create distinguishable results for a request you control. The same applies to REJECT: a connection error doesn’t prove the reject rule worked. A request that failed DNS resolution may never have entered the expected branch.

Treat every move as a change in scope

When reviewing an existing config, mark the fallback first, then work upward from the bottom to find broad rules, especially keyword rules, suffixes, and large IP prefixes. For each new exception, identify the existing rules it overlaps with and place it before the earliest overlap. If an old rule is a general condition that follows other exceptions, don’t move the whole block to the top; that could mask existing reject or direct rules. Keep one representative request from the original setup as a comparison to quickly catch unexpected changes caused by reordering.

When merging a configuration snippet from someone else, first verify its section, policies, and prerequisites. Don’t paste in a second [Rule] header, another file’s default rules, or a policy name that hasn’t been defined. For configs that use remote rule sources, distinguish the visible local order from content changes caused by source updates. If a hostname that never matched before is suddenly caught by an earlier set, adding another exception at the end won’t explain the change. Record the actual config source and test copy used when retesting.

When troubleshooting rule priority, move or edit only one rule per round, and keep a request that should remain unaffected as a comparison. Confirm the result before moving on to the next rule.

Common Patterns: Choose the Smallest Rule Set That Fits

Proxy only known targets

[Rule]
DOMAIN,signin.example.com,DIRECT
DOMAIN-SUFFIX,example.com,PROXY
FINAL,DIRECT

These rules keep the login host direct, send other hosts with the same suffix through a proxy, and leave all remaining requests direct. This works when the scope is clear and can be expressed with a small set of hostnames. Test the login host, one business host under the same suffix, and a target outside that suffix. If the service later moves its file hosting to a different suffix, the existing rules won’t cover it automatically. Verify the new hostname and access requirements, then add a narrow rule instead of changing FINAL just to restore one feature.

If a service opens before login but only some features fail afterward, the problem may occur when a request moves from the login host back to the service host. To see whether different paths are involved, list the hosts used by the page, authentication, and API requests during one complete operation, then check each policy. If all these hosts need the same path, base the routing rules on the service’s dependencies. Don’t assign one policy to every hostname that happens to contain the same word. Follow the access method required by the service provider.

Use a proxy fallback while keeping explicit exceptions for local services

[Rule]
DOMAIN,printer.example.net,DIRECT
IP-CIDR,192.0.2.10/32,DIRECT,no-resolve
FINAL,PROXY

This example sends connections to a specific service by name direct, and also sends direct connections to a specified address direct. These are documentation values; they don’t mean the device can use them to find a local service. In a real setup, first check that the chosen name resolves correctly and that the current network has a route to the device. Then choose a domain rule, an IP rule, or both to match the actual setup. If the device’s address is assigned dynamically, a single-address rule needs maintenance. A broad address range may also exempt connections unrelated to the service.

Using the same configuration on two networks doesn’t mean the same private address points to the same device. An address that reaches an office device at work may reach a different device—or no device at all—at home. Include the network environment in your test expectations: verify the target’s identity and direct access on the original network, then confirm that another network won’t unexpectedly reach a different service at the same address. An IP rule controls how a request is sent; it doesn’t verify the remote service’s identity.

Keep the service host available; reject an independent request precisely

[Rule]
DOMAIN,stats.example.org,REJECT
DOMAIN-SUFFIX,example.org,DIRECT
FINAL,PROXY

When you’ve confirmed that a separate host isn’t needed, use an exact reject rule. Before changing anything, list the actions that must keep working, such as loading the main page and saving settings. Repeat those steps after adding the rule. If a critical feature shares the host you want to reject, this rule can’t distinguish between them. If the entire suffix should remain direct, the reject rule’s position matters: placing it after the suffix rule means the earlier direct rule will take precedence and the request won’t be blocked.

For routine maintenance, leave a brief note explaining why a rule is needed, which target it depends on, and which line to revert if it fails. When sharing a snippet, include the surrounding rules that affect its scope, especially rather than copying just FINAL or a keyword rule that looks plausible on its own. Before adding an example to an existing config, compare it with the current fallback. If you only need one exception, don’t also change the behavior of unlisted targets.

Testing and Rollback: Narrow Down the Failure Step by Step

Establish a repeatable baseline first

Before you start, keep a copy of the most recently working configuration. Check the selected file in Config, and note the network, mode, and test targets. Record the results for one target known to work direct and another that normally requires an existing server. Check server selection in Home and use Connectivity Test as a reference, but don’t assume one successful test means every target is reachable. If both paths fail, check the device’s network first. If only the proxy path fails, verify the existing server address, authentication details, and protocol settings.

If Config fails but both comparison paths work, note which rule should match the target in the current config, its policy, and every earlier condition that could catch it. Start a new connection, repeat the same action, and check the diagnostic records available on your platform. If the fallback matched, check the hostname or address actually visible to the client. If an unexpected general rule matched, check the order. If the policy is correct, investigate DNS, the connection, or the target service instead of adding more match rules.

Troubleshoot by failure stage; change one variable at a time

  1. Configuration won’t load: Go back to the last change and check the section, punctuation, number of parameters, and policy spelling. In the same configuration copy, undo only that change and see whether it loads again.
  2. Can’t resolve the target address: Check DNS settings and whether [Host] contains an outdated mapping. If only one network fails, keep the other network’s DNS result as a comparison; don’t change servers at the same time.
  3. The target has an address, but the connection fails: Check the current mode, the matched policy, and the connection path. Check whether a change in address family sent the request down a different branch, or whether the expected network is unavailable.
  4. The connection succeeds, but the service is incomplete: Find follow-up requests triggered by the failing action. Check whether they need another host, whether the redirect destination is reachable, and whether a reject rule is too broad.

For example, if the system resolves two addresses but only one path fails, note which address the request used and check whether the IP rule covers only one address family. If a certificate error occurs when accessing the same domain, check whether a static mapping points to the wrong server. Don’t bypass certificate errors just to pass a test. Identifying the failure stage preserves useful clues; repeatedly changing the mode, DNS, and server makes the results harder to compare.

[Rule]
DOMAIN,check.example.com,PROXY
FINAL,DIRECT

This minimal rule set separates one proxy target you control from all other direct targets. Replace the domain with your actual test host and verify it in a separate copy, keeping any network settings the original config needs. Don’t delete a working configuration, clear existing server details, or rewrite a subscription you use day to day just to create a comparison. If the minimal setup works, add back domain exceptions, address classifications, and rewrite rules one section at a time. The first section that reproduces the issue is where to focus next.

Retest after rollback

If you can’t explain the scope of a new configuration, reselect the backup, restore the required Global Routing mode, and retest the original scenario and one comparison target that should work. The disappearance of an error page alone doesn’t confirm a successful rollback. If you need help, share sanitized rule lines, the expected policy, network type, and steps to reproduce. Don’t publish a full subscription URL, authentication details, request URLs with sensitive parameters, or unredacted logs. For more troubleshooting paths, see Troubleshooting and Why Shadowrocket Is Connected but Pages Won’t Load: A Step-by-Step Guide.

To switch devices, choose the right option on the all-platform download page. Get Shadowrocket for iPhone and iPad from the App Store product page for a one-time US$2.99 purchase; check the App Store page for system requirements. The Mac (Apple silicon and Intel), Windows, Android, and Linux versions are free, with installers available only from this site’s /dl/ path. A one-time app purchase is not a service plan; on a new device, you’ll still need your existing subscription or server details.