03Open-source ecosystem
Project V, V2Fly, and Xray
Knowing which names refer to projects, cores, or graphical clients explains why a guide may not apply word for word to every app—and helps narrow down troubleshooting.
- Client layer
- Subscriptions, node lists, configuration management
- Core layer
- Inbound, routing, outbound
- System layer
- App integration, network permissions
Project V's inbound and outbound model
Project V is a family of projects for proxy tools, with V2Ray as the core component that handles connections. The V2Fly community later continued V2Ray development, while Xray split from this codebase and evolved independently. They share several basic concepts: inbound receives traffic, routing determines where it goes, and outbound hands the connection to the next hop or rejects it according to the configuration. The four outbound names shown above are illustrative labels; actual names and behavior depend on the configuration.
For example, a browser request reaches the relevant inbound only when the browser uses the local proxy settings. The core then matches rules using the destination information available to it. If the browser bypasses the proxy, the core won't route its traffic even if it's running. Understanding this boundary is more useful than memorizing where a button was in a particular version: settings can change, but traffic still has to enter the core first.
Core branches and configuration compatibility
V2Fly and Xray maintain their own protocol implementations and configuration features. Their documentation may include fields such as inbound and outbound, but fields with the same name aren't necessarily interchangeable in every option. VMess and VLESS describe parts of a protocol; a connection also depends on the transport, security settings, and how both ends are implemented. The server configuration determines what the client needs to support—don't infer that from a region or route name in the node description.
When migrating, keep the old setup and import the same known-good configuration into the new environment on its own. If the UI recognizes the node but the core won't start, check the startup error first. If the core starts but that node won't connect, check the node parameters and network reachability. Following this order helps distinguish mismatched transport settings from an expired subscription and avoids changing all your routing rules up front.
Open-source status of v2rayN, v2rayNG, and v2flyNG
All three clients make their source code public. v2rayN is a desktop client with tools for managing subscriptions, nodes, and routing settings. v2rayNG uses the Xray core, while v2flyNG offers a mobile option based on the V2Fly core. Clients and cores are separate project layers: how subscriptions are displayed, where menus appear, or how system permissions are requested doesn't tell you whether two cores belong to the same branch.
Open-source licenses govern how code can be used, modified, and redistributed; they don't tell you whether to trust a third-party subscription service. Using a client for free and paying for a network service are separate matters. Don't post an entire configuration file publicly if it contains subscription URLs or node credentials. When reporting an issue, you usually only need to include the operating system, client and core names, the stage where the failure occurs, and an error summary with sensitive fields removed.
Client updates, core updates, and subscription refreshes
A client update may change the interface or how it integrates with the system. A core update may affect support for new features or how configurations are parsed. A subscription refresh retrieves the node list again. Installers sometimes bundle a core, so one upgrade can affect both the client and core layers. Refreshing a subscription, however, doesn't update the app or automatically fix local routing errors. When troubleshooting after an update, first find out exactly what changed.
After saving your old settings, compare before and after using the same network, node, and test destination. First check that the app launches, then that the node still works, and finally whether the existing rules match as expected. If only the routing result changed, also check whether domain or IP data was updated as part of the upgrade. Making one change at a time gives you a clear rollback point and makes it easier to provide maintainers with a useful reproduction.
Testing connections and isolating issues
Start by following the subscription import guide to get a working setup before trying more complex rules. If the list doesn't update, check the subscription source and retrieval process. If the list is current but every node fails, check your network, connection mode, and whether the core started. If only certain apps fail, check whether they use the system proxy or their own DNS settings, and which rule matches their destinations.
Don't rely on a single latency result to compare connection quality. Short web requests highlight response time, while file transfers better reflect sustained throughput; different factors affect each. Repeat comparisons on the same network and with the same test destination, noting any network, node, or configuration changes. If requests remain unstable after reverting settings, check the node and route again rather than adding more local rules.