Shadowrocket Glossary
Look up interface terms, protocols, and rule syntax. First distinguish when to connect, where to connect, and how to route traffic, then adjust the relevant settings.
Interface labels and rule keywords remain in English on this page; the explanations describe their purpose without replacing the names used in the app. Menu locations may vary by platform. For step-by-step instructions, see the tutorial; for complete syntax, see the rule reference.
Interface and Toggles
When you find a setting, first determine whether it manages connections, changes the routing mode, or displays diagnostic information. Similar names do not mean the settings work the same way.
- Home
- The screen for viewing servers, selecting a connection target, and controlling the connection toggle. A saved entry only means its settings are stored; it does not mean the connection is active. When troubleshooting, check the selected server, connection toggle, and Global Routing mode separately rather than relying on just one.
- Global Routing
- Controls the overall routing mode. Common options are Config, Proxy, and Direct. Config follows the rules in the current configuration; Proxy routes traffic through a proxy; and Direct connects directly, subject to necessary system and app behavior. The interface modes Proxy and Direct differ from the uppercase PROXY and DIRECT policies in rule entries. To test routing rules, select Config first.
- On Demand
- Triggers connection behavior based on preset network conditions, such as starting a connection on a specified network. It controls when to connect, not which policy a domain uses. Verify that a manual connection works before enabling it, then configure the trigger conditions one at a time. Otherwise, automatic connections can obscure server or configuration issues.
- Add Server
- The option for entering existing server details manually. Type selects the connection protocol. The server address, port, authentication fields, and transport options must match the server settings; a name alone is not enough to identify the protocol. Get the parameters from your provider. If you already have complete import details, check the imported result rather than creating a duplicate entry.
- Connectivity Test
- A diagnostic tool that checks whether a server responds using the selected test method. Latency usually measures the time for one test interaction; it is not available bandwidth and does not directly indicate the speed of every website. For meaningful comparisons, keep the test method and network consistent. Even after a successful test, verify DNS, routing, and application-layer connectivity with an actual request to the destination.
- Data
- A screen showing traffic readings such as uploads and downloads. Use the views available in the client to review server or app usage. Readings cover a specific statistical range and should not be treated as a provider’s bill; resets and reporting periods also affect comparisons. To verify a rule, check request records and the matched policy. An increase in total traffic alone does not confirm that a particular rule worked.
- DNS
- The process of translating a domain name into an IP address, usually one of the steps before connecting to a domain. Successful resolution only means an address was returned; it does not guarantee that the service at that address is available. Conversely, a DNS failure can make a working server appear unreachable. Before changing DNS, note the original settings and distinguish resolution errors from connection timeouts and certificate errors.
Protocols
A protocol name describes how a connection works, not its quality. Choose Type based on the server details you already have; using the same address with a different protocol does not automatically produce a working configuration.
- Shadowsocks
- A proxy protocol that establishes a connection using a specified encryption method and password. The address, port, encryption method, and password must all match, along with any additional parameters required by the server. If authentication or the handshake fails, first check for truncated fields or extra spaces. Changing rule policies will not fix incorrect protocol settings.
- VMess
- A proxy protocol that authenticates using parameters such as a user ID. The connection may also have separate transport and TLS settings. After entering the user ID, check the transport type, path, server name, and other related fields. The port number alone does not reveal these settings. When troubleshooting, also check that the device clock is accurate, since a time difference can affect authentication.
- VLESS
- A proxy protocol that handles user authentication separately from transport security settings, with the user ID as a key parameter. The server determines the required transport type and security options; do not copy them from another server entry. If the user ID is correct but the connection still fails, check TLS, the server name, and transport fields rather than repeatedly changing the authentication details.
- Trojan
- A proxy protocol that typically uses TLS and password authentication. The password, server address, port, and server name used for certificate verification must meet the server’s requirements. Certificate errors and password errors occur at different stages, so check them separately. Do not disable certificate verification or ignore server identity just to make a test pass temporarily.
- HTTP
- In the server type list, HTTP means connecting to an HTTP proxy server; it does not mean the destination website can only use HTTP. If supported by the proxy, CONNECT can establish a connection to the destination. Whether a username and password are required, and whether TLS is used between the client and proxy, depends on the server settings. This is separate from whether the destination website uses HTTP or HTTPS.
- SOCKS5
- A protocol for sending connection requests to a proxy server. The destination can be specified by domain name or address. TCP, UDP, and authentication support depend on both the client and server; success with one test does not prove that every traffic type will work. SOCKS5 is not itself an encrypted tunnel, so consider the underlying transport environment when using it.
- WireGuard
- A protocol that creates an encrypted tunnel using key pairs. Its parameters typically include the Endpoint, keys for both sides, tunnel addresses, and allowed address ranges. This is not a standard server entry that works by simply replacing a password with a key; the fields must correspond to one another. A private key is sensitive information. Remove it before sharing diagnostic screenshots or configuration files, and do not swap the private and public keys.
- Hysteria2
- A QUIC-based proxy protocol that relies on UDP and uses TLS with corresponding authentication parameters. If the current network restricts UDP, the handshake may time out even when the address and authentication details are correct. First check the network conditions and server reachability, then verify authentication and other options. The protocol name alone does not guarantee a particular speed.
Rule Keywords
Rule keywords define the conditions used for matching. The snippets below illustrate syntax; they are not complete configurations. example.com and its subdomains are included only to demonstrate matching boundaries.
- DOMAIN
- Matches one exact domain and does not automatically include its subdomains. For example,
DOMAIN,example.com,PROXYmatches example.com but not api.example.com. Use it to set a specific exception for one host. To cover a domain and its subdomains, consider DOMAIN-SUFFIX and check the order of any exception rules.
- DOMAIN-SUFFIX
- Matches a specified domain and its subdomains. For example,
DOMAIN-SUFFIX,example.com,PROXYmatches both example.com and api.example.com. It respects domain boundaries rather than matching any string ending, so it does not cover notexample.com. To use a different policy for a particular subdomain, put a more specific DOMAIN rule first.
- DOMAIN-KEYWORD
- Checks whether a domain contains a specified string, for example,
DOMAIN-KEYWORD,ads,REJECT. This matching method has a broad scope and may include domains you did not intend. Unlike DOMAIN-SUFFIX, it does not respect domain boundaries. After adding the rule, review actual match records. If it catches too much, use a specific domain or domain suffix instead.
- GEOIP
- Matches the destination IP’s geolocation according to IP geolocation database records, for example,
GEOIP,CN,DIRECT. It does not identify a website’s language or directly indicate where its operator is based. Database updates and the destination’s resolved IP can both affect matches. Understand the relevant DNS behavior before using it, and do not treat geolocation as the sole way to identify a destination.
- IP-CIDR
- Matches IPv4 address ranges. For example,
IP-CIDR,192.168.0.0/16,DIRECTcan target a range of private addresses. The number after the slash is the prefix length, which determines the range; it is not a port number. Whether the rule triggers domain resolution also depends on configuration options. When troubleshooting local network access, check the actual subnet, destination address, and rule position.
- IP-CIDR6
- Matches IPv6 address ranges. For example,
IP-CIDR6,fc00::/7,DIRECTrepresents a range of IPv6 unique local addresses. An IPv4 IP-CIDR rule does not automatically cover IPv6 addresses, and the same domain may return both address types. If requests take different routes, check the actual connection address rather than inferring the result from the domain alone.
- USER-AGENT
- Matches requests based on the User-Agent in visible HTTP requests; it is not a reliable way to identify the app that made the request. Encrypted connections may not expose this field, and apps can omit or change it or share the same identifier. Before using this type of rule, confirm that the request details are visible and understand the certificate and privacy implications of HTTPS decryption. Do not base a match solely on an app’s name.
- FINAL
- A fallback rule used when none of the preceding rules match, for example,
FINAL,PROXY. It usually goes at the end of the Rule section and has no additional domain or address condition. If FINAL appears too early, ordinary rules below it will never be checked. Choose a fallback policy based on whether uncategorized requests should connect directly or use the proxy route.
Policies and Match Order
Keywords define the match conditions; policies define what happens when a rule matches. More specific conditions are not automatically placed first, so the order you write them in is part of the configuration.
- PROXY
- A rule policy that sends matching traffic through the proxy route. The actual server depends on the current selection and related settings. PROXY is not a protocol and does not mean traffic always uses a particular server address. If a request matches PROXY but fails, check the selected server, authentication parameters, and destination reachability rather than assuming the rule did not match.
- DIRECT
- A rule policy that connects matching traffic directly to its destination without using the selected proxy server. Direct connections still depend on the current network’s DNS, system routing, and destination service, so DIRECT does not guarantee a successful connection. It differs from the Direct mode in Global Routing: a DIRECT rule applies only to requests that match that rule.
- REJECT
- A rule policy that rejects matching requests. It does not delete server entries or clear subscriptions. An overly broad match can also reject requests an app needs, causing missing content or unavailable features. When this happens, check the match records first, narrow the conditions, or add a specific exception. Avoid simply piling on more reject rules.
- Match Order
- Ordinary rules in Rule are checked from top to bottom. The policy for the first match is applied, and later conditions are not checked just because they seem more specific. Put specific exceptions before broad rules and place FINAL at the end. After making changes, confirm that the configuration is active and Global Routing is set to Config, then check the match result with a new connection request.
Subscriptions and Configuration Files
Server details answer “where to connect”; rules answer “which requests use that connection.” Importing servers, updating a subscription, and switching configurations are separate actions and should be verified separately.
- Subscribe
- A way to retrieve and update a group of server details using a subscription URL, which you get from your provider. The URL may contain access credentials; do not share it in screenshots, logs, or shared configurations. Updating a subscription can change the server list, but does not switch the rule configuration or guarantee that the previously selected server is still available. After updating, check the current selection again.
- Server
- A saved connection entry in the client, usually containing a protocol, address, port, and authentication parameters. It is also commonly called a node. The entry name is just a label for identification; it does not confirm the server’s location, protocol capabilities, or connection quality. Servers can be added manually or imported from an existing subscription. Importing the same details more than once adds management overhead without improving the connection.
- Config
- Can refer either to the place for managing configuration files or to the Global Routing mode that applies configuration rules to traffic; its meaning depends on context. A configuration file can include sections such as General, Rule, and Host for basic behavior, matching conditions, and static domain mappings. Saving a file does not mean it is currently in use. Keep a copy before making changes, then check which configuration is active.
- URL Rewrite
- A configuration section for rewriting request URLs or applying related rewrite actions based on conditions. It differs from Rule routing policies such as PROXY, DIRECT, and REJECT. A rewrite can change the request path or response flow; it cannot replace server authentication parameters. HTTPS requests are also subject to visibility and certificate settings. Understand the match scope first, then test each rule individually.