PROJECT V / CORE LINEAGE
The Project V ecosystem and open-source clients
The client, core, and server configuration are three distinct layers. The client handles importing, switching, and system integration; the core handles protocols, resolution, and forwarding; the server configuration determines the parameters used at both ends of the connection. Switching graphical clients doesn’t change a node’s protocol or convert the existing server settings.
From Project V to independently developed cores
The Project V ecosystem grew around modular network proxy tools. V2Ray handles core traffic forwarding, with configurations separating inbound, outbound, and routing so local traffic reception can be configured independently of the remote connection. The V2Fly community continues V2Ray’s maintenance and development; Xray evolved from the existing V2Ray core and is maintained on its own schedule. The projects share some configuration concepts, but they’re released independently by different communities. Updates from one project don’t determine how the other behaves.
This distinction is useful when troubleshooting: if you can’t find a button, check the graphical client’s interface guide. If the logs report an unrecognized field, check whether the core supports it. If both sides can read the configuration but the connection still fails, verify the server address, transport, and security parameters.
A protocol name isn’t a complete configuration
Open-source projects publish protocol documentation and implementations that can help explain why the same configuration behaves differently across cores. VMess, VLESS, and Trojan each have distinct configuration requirements. REALITY handles transport security; it isn’t an interchangeable alternative to these proxy protocols. When choosing a client, first identify the protocols, transport, and security mechanisms required by your configuration, then confirm that the client and core support those fields together.
When migrating, distinguish between “the link imports” and “the connection works.” The first means the client can at least parse the input; the second also depends on core behavior, parameters on both ends, and the current network. If you encounter an unfamiliar field, start with the concept reference to identify its layer, then decide whether to check the client settings or the core documentation.
What each client is for
v2rayN, v2rayNG, and v2flyNG are all open-source graphical clients. v2rayN is designed for desktop and provides subscription management, a local proxy, and system proxy controls; supported cores and modes depend on the specific build and configuration. On Android, v2rayNG uses the Xray core and Android’s VpnService to manage which traffic is routed. v2flyNG is an Android option based on the V2Fly core, suited to setups that need to match V2Fly configuration and behavior.
Before switching between the two Android clients, check where your configuration came from and which core it requires, then note your per-app list and routing preferences. These settings are stored locally in the client; saving the subscription URL alone won’t restore them. If both clients are installed, disconnect the active VPN before starting the other one so you can tell which app is routing your traffic.
Client, core, and subscription updates are separate
A client update can move menus or affect system compatibility; a core update can add fields or change defaults; a subscription update rereads configuration entries from its source. These are independent actions. If a node disappears after updating a subscription, check the list returned by the source first. If a configuration error starts after updating the client, inspect the core and the field causing the problem instead of blindly overwriting the existing group.
Before making changes, note the group, connection mode, and key routing rules for a configuration that currently works. After updating, test that configuration before trying new features. If you need to investigate further, record whether the failure occurs during import, startup, or a request. Before sharing logs, hide account identifiers, full subscription URLs, and node credentials; keep only the error details needed to identify the problem.