In Shadowrocket’s Add Server, VMess and VLESS aren’t interchangeable speed settings. The Type must match the provider’s server configuration; then check the UUID, port, transport, and TLS. This guide is for users who already have a subscription or connection details from their provider and need to verify a manual setup. It explains how the two protocols handle authentication and encryption, and which fields to check when a port is reachable but webpages won’t load.
1. Start with the protocol-level differences between VMess and VLESS
VMess authenticates with a user ID and can encrypt payloads at the protocol level. VLESS uses a leaner authentication and forwarding design and does not encrypt payloads on its own, so you need to consider the outer security layer to understand how the full connection is protected. That means “VLESS doesn’t encrypt by itself” does not mean every VLESS connection sends data in plaintext, nor does it mean you can skip TLS settings.
Both protocols may use a UUID as the user ID, but sharing a UUID doesn’t mean the server accepts both protocols. The server’s listening protocol, user credentials, and transport settings must all match as a complete set. Changing Type from VMess to VLESS only changes how the client sends data; it does not switch the server to VLESS.
| What to check | VMess | VLESS |
|---|---|---|
| User authentication | Usually uses a UUID and a protocol-specific authentication mechanism | Usually uses a UUID, with an authentication design different from VMess |
| Payload encryption at the protocol level | May be provided; follow the server parameters | Not provided by VLESS itself; check the outer security settings |
| Protocol-specific parameters | May include encryption options and the legacy compatibility field AlterId | Some configurations may include Flow; don’t copy VMess fields across |
| Transport and security layers | Can be used with TCP, WebSocket, TLS, and more | Can also be used with TCP, WebSocket, TLS, and more; settings must match the server |
| When to choose it | Your provider explicitly supplies a VMess configuration | Your provider explicitly supplies a VLESS configuration |
Modern VMess deployments generally use AEAD authentication; don’t reuse AlterId values from older guides. Modern configurations from providers commonly specify AlterId as 0, but always follow the original parameters. A device clock that is significantly off can also cause VMess authentication to fail, so check that automatic date and time are enabled in system settings.
Key point: choose the Type first, then verify the full connection setup
If your provider gives you VMess, keep VMess; if they give you VLESS, keep VLESS. To test the other protocol, get its complete configuration rather than changing Type and reconnecting repeatedly.
2. Fields in Add Server you must not mix up
In Home, tap the add control to open Add Server. Choose Type first, then fill in the other fields. Layouts may vary across platforms and app versions, and available fields depend on the Type and transport options you select. If a field isn’t shown, don’t put its value in Password, Host, or the remarks field instead.
- Server address and port: Enter the server’s domain name or IP address in the address field, and enter the port separately. For example,
edge.example.comand443belong in different fields. Don’t puthttps://, a path, and the port all in the address field. - UUID: Copy the complete user ID, keep the hyphens, and remove any spaces at either end. A UUID is not a subscription URL token or your account password.
- VMess encryption option: Enter the value supplied by your provider. Don’t assume VLESS’s
encryption=nonemeans encryption should also be disabled for VMess. - Transport: TCP and WebSocket are different options. If you choose WebSocket, also check the Path and any Host value.
- TLS: Check the toggle, server name, and certificate verification requirements. Port
443does not automatically mean TLS is configured correctly. - Flow: Set this to the supplied value only when your provider requires it and the current interface supports the setting. Don’t add a Flow value from another guide if your VLESS configuration doesn’t specify one.
Shadowsocks, Trojan, Hysteria2, and WireGuard each have their own authentication fields and transport mechanisms. Even if they use the same server address, you can’t build a connection by filling in a VMess or VLESS form. Add Server’s Type selects the protocol; it isn’t a display name.
3. How TLS, WebSocket, and protocol Type work together
When checking connection details, think in three layers: the inner protocol is VMess or VLESS; the transport may be TCP or WebSocket; and the outer layer may use TLS. Each setting serves a different purpose. A correct UUID alone doesn’t prove the entire connection is configured properly.
How to check two connection setups
VMess + WebSocket + TLS
- Set Type to VMess, then enter the UUID and encryption option from the original parameters.
- Choose WebSocket for transport and check the Path character by character, such as
/ws. - Check the WebSocket Host and TLS server name separately.
- Make sure TLS is enabled and the port matches the one the server is listening on.
VLESS + TCP + TLS
- Set Type to VLESS and enter the corresponding UUID.
- Choose TCP for transport; don’t add a WebSocket Path.
- Enable TLS and enter the server name supplied by your provider.
- Enter Flow only when explicitly required, and check that the client offers the corresponding option.
These examples illustrate which fields to check; they are not ready-to-use connection settings. Use the complete configuration from your provider for the actual address, port, path, and security parameters.
The server address determines where the connection goes. The TLS server name is used during the handshake and certificate name verification. WebSocket Host is part of the HTTP request and may be used for reverse-proxy routing. These values may match, but they can also differ. If you enter an IP address in the address field, don’t automatically replace the TLS server name with that IP.
Take the example path /ws: /WS and /ws/ may not be equivalent. If the upstream service matches paths exactly, a single-character difference can cause an HTTP 404 or a failed WebSocket upgrade. If the certificate name doesn’t match, first check the server name, system time, and provider settings; don’t disable certificate verification as a workaround.
- TCP connection times out: Start by checking the address, port, local network, and server reachability.
- TLS handshake error: Start by checking the TLS toggle, server name, certificate, and system time.
- HTTP 404 or upgrade failure: For WebSocket connections, check Host, Path, and the server’s forwarding configuration.
- Traffic still won’t pass after the handshake: Recheck the UUID, protocol Type, and other authentication parameters.
4. With an existing subscription, keep the imported configuration or rebuild it manually?
If you already have a working subscription, keep the complete imported settings wherever possible. Get subscriptions from your provider. Importing usually brings in the protocol, port, transport, TLS, and other fields together, making it easier to keep them consistent than filling them in manually based on a server name alone. A node remark saying “VLESS” or “VMess” is only a clue; open the edit screen and check Type to confirm.
Keep the imported subscription settings
RecommendedIn Home, import or update your existing subscription first, then open the target entry’s edit screen to check Type, port, and TLS. Don’t change fields before testing, so you can refer back to the original settings.
Best for: Checking the protocol Type and connection settings in an existing provider subscription
Rebuild the settings manually in Add Server
Once you have the complete parameters, create a separate entry and fill in Type, address, port, UUID, transport, and TLS in that order. Keep the original entry so you can compare settings and restore them if needed.
Best for: When your provider supplies field-by-field instructions or you need to pinpoint a setting that differs
A subscription URL and a single-node share link are different kinds of input. https://example.com/sub?token=xxxx is an illustrative subscription URL, not a working connection. vmess:// and vless:// are common protocol prefixes for single-node share links. Complete share links contain credentials and shouldn’t be posted publicly.
- Before importing, make sure you copied the entire link without truncating query parameters or including explanatory text or line breaks.
- After importing, open the entry’s edit screen and check that Type matches your provider’s instructions, then verify the transport and TLS settings.
- If you need to make changes, create and clearly name a separate test entry. This makes it easier to tell the original settings apart from your manual changes after a subscription update.
- A subscription update may overwrite settings on managed entries. When you finish testing, check the original entry again so you don’t mistake a temporary change for a permanent configuration.
5. Troubleshoot connection failures in order: network, handshake, then routing
Turning on the connection toggle in Home means the client has started connecting; it doesn’t guarantee that every website is reachable. A response from a port test only shows that part of the tested route is reachable. It doesn’t confirm that VMess or VLESS authentication, TLS, and app requests are all working.
- Check a direct connection: First make sure your Wi-Fi or mobile network can reach a known-accessible website without a proxy. If direct access also fails, troubleshoot the local network before changing protocols.
- Check the original settings: In Home, open the target entry and compare Type, address, port, UUID, transport, and TLS in that order. Change only one setting at a time.
- Temporarily isolate routing rules: Select the target entry and temporarily switch Global Routing to Proxy for a test. If Proxy works but Config doesn’t, check the rules and policies first instead of changing Type.
- Check which rules match: In Config, inspect the active configuration file.
DOMAIN-SUFFIXmatches domain suffixes,IP-CIDRmatches IP ranges,GEOIPmatches by IP geolocation, andFINALhandles requests that didn’t match earlier rules. Rule order and target policies both affect the result. - Restore your usual settings: After testing, restore your original Global Routing setting. If Settings → On Demand has automatic connection rules configured, note their status so automatic reconnections don’t interfere with manual disconnect and reconnect tests.
In Global Routing, Proxy routes traffic through a proxy, Direct connects without one, Config applies the selected configuration file, and Scene switches according to configured scenarios. None of these changes the server-side protocol from VMess to VLESS. To isolate a protocol issue, keep the same entry, network, and destination, and avoid changing several conditions at once.
Key point: identify where the failure occurs, then adjust the relevant setting
Record “network timeout,” “TLS handshake failure,” “incorrect WebSocket path,” and “wrong routing policy” separately. If a destination works with Proxy but not Config, check which Config rule matched first. Don’t try to fix a routing issue by changing the UUID or Type.
6. Common questions about VMess and VLESS settings
My provider gave me VMess. Will switching to VLESS make it faster?
Don’t switch it directly. Keep the original VMess entry. If your provider also offers a VLESS configuration, create a separate entry for it. Compare them on the same network, at similar times, and against the same destination; check latency, sustained transfers, and failures. The protocol name alone doesn’t tell you which will be faster.
VLESS says none. Should I turn off TLS?
No. VLESS’s protocol-level encryption setting and outer TLS are separate. Open the target entry’s edit screen and check the transport and TLS settings against your provider’s instructions. If the original configuration requires TLS, leave it enabled and use the correct server name.
Port 443 is reachable. Why won’t webpages load?
A reachable port doesn’t mean application-level authentication succeeded. First check Type and UUID, then the TLS server name. If you’re using WebSocket, also check Host and Path. Once the fields match, briefly test with Global Routing → Proxy to distinguish a protocol issue from a routing issue.
Why did my manual settings change after I updated the subscription?
An entry managed by a subscription may be reset to the provider’s settings during an update. If you need to keep testing, create a separate manual entry and leave the original subscription entry intact. If the entry remains unusable after an update, tell your provider the entry name, where the error occurs, and which fields you checked; don’t share the full subscription URL publicly.
VMess suddenly stopped working. Should I switch protocols first?
First check that automatic date and time are enabled, then confirm whether the subscription was updated or your provider changed any parameters. Test on another local network for comparison. Only try another protocol after you have its complete configuration; a single timeout doesn’t mean VMess itself has stopped working.