-
Notifications
You must be signed in to change notification settings - Fork 2
Client Configuration
English | 简体中文
Configure both endpoints against the same centre identity. A server-side deployment is not complete until real clients register and transfer a desktop session.
Open the RustDesk client Settings → Network, unlock the settings, and use:
| Client field | Value |
|---|---|
| ID Server | Public HBBS hostname, for example id.example.com, or an explicit :21116 address when non-default. |
| Key | Complete one-line content of the centre id_ed25519.pub. This is not an API token or licence key. |
| Relay Server | Leave empty when Starry HBBS should dynamically allocate a Geo Relay. |
| API Server | Optional separate HTTPS URL, such as https://api.example.com, only when an API is deployed. |
| Use WebSocket | Off by default; enable per client only when that network needs WSS. |
The official RustDesk client configuration guide also distinguishes ID, Relay, API, and public Key fields. Starry does not change their meaning.
Native clients require both 21116/TCP and 21116/UDP. With the official
1.1.16 server, controlled-endpoint registration and heartbeats use UDP while
controller initiation uses TCP/Secure TCP; disable-udp does not turn native
registration into a TCP-only path. Controlled endpoints that must disable UDP
should use WSS with complete /ws/id and /ws/relay deployment.
A non-empty static Relay field tells the client to use that address and can bypass the dynamic Relay address returned by HBBS. Leave it empty to exercise Starry rules and failover.
If you temporarily set a Relay for diagnosis, remove it before validating Geo allocation.
| Client A | Client B | Signalling/data expectation |
|---|---|---|
| Off | Off | Native signalling. P2P preferred; native HBBR only when required. |
| On | Off | A uses WSS, B remains native. The session is Relay-only through one HBBR using mixed transports. |
| Off | On | Same as above with directions reversed. |
| On | On | WSS signalling and WSS Relay; no normal P2P path. |
The server setting websocket_signal.enabled permits the path but never
changes a client switch. Conversely, a client switch cannot compensate for a
missing /ws/id, /ws/relay, or invalid certificate.
An API is optional for base self-hosted control and is not supplied by Starry.
A compatible third-party API can be used; the recommended integration is
q1ngyang/rustdesk-api-kessoku.
When an API is used:
- verify the HTTPS API status and login independently;
- confirm the API exposes the same HBBS public key and ID Server expected by the clients;
- leave Relay Server empty for Geo allocation; and
- repeat the remote-control test while logged in, because authenticated
clients may negotiate Secure TCP on
21116/TCPbefore punch-hole or Relay.
Successful API login is not evidence that Secure TCP, HBBR, or the desktop data path works.
See Account/API Integration before adding accounts or connection tokens.
Use two known devices and record:
- client versions and platforms;
- both configured ID/API/Key values, with secrets redacted;
- public egress addresses as observed by HBBS;
- whether each WebSocket switch is on;
- whether the final session is P2P or Relay;
- selected Relay hostname and shared Relay UUID; and
- timestamps for matching client/server logs.
Test native first, then API login, then WSS. Change one dimension at a time.
- Copying
id_ed25519instead of the public.pubcontent. - Treating an API access token or a commercial licence key as the server Key.
- Leaving an old static Relay address in one client.
- Enabling WebSocket on only the client while the server paths are incomplete.
- Assuming WSS should preserve normal P2P; it is intentionally Relay-only.
- Testing two clients behind one NAT without using the same observed public IP
twice in
test-geo. - Comparing logs from different attempts instead of one timestamped session.
Continue with Operations and Verification.
- Documentation home
- Getting started
- Docker image usage
- Docker deployment
- Native deployment
- Multi-node deployment
- Reverse proxy and TLS
- Client configuration
- Account/API integration
- Configuration reference
- Connection authentication
- Control Agent
- GEO rules: basics
- GEO rules: advanced
- Operations and verification
- Troubleshooting
- Upgrade and rollback
- Architecture and build
- 文档主页
- 快速开始
- Docker 镜像使用
- Docker 部署
- 原生部署
- 中心与 Relay 多节点部署
- 反向代理与 TLS
- 客户端配置
- 账户与 API 服务接入
- 配置参数详解
- 连接认证
- Control Agent
- Geo 规则:入门
- Geo 规则:进阶
- 运维与完整验证
- 常见问题排查
- 版本升级与回滚
- 架构与构建