Skip to main content
The Gateway is fully supported on Linux. Node is the recommended runtime; Bun is not recommended (known WhatsApp/Telegram issues). There is no native Linux companion app yet. Contributions are welcome.

​
Quick path (VPS)

  1. Install Node 24 (recommended) or Node 22.19+ (LTS, still supported).
  2. npm i -g openclaw@latest
  3. openclaw onboard --install-daemon
  4. From your laptop: ssh -N -L 18789:127.0.0.1:18789 <user>@<host>
  5. Open http://127.0.0.1:18789/ and authenticate with the configured shared secret (token by default; password if gateway.auth.mode is "password").
Full server guide: Linux Server. Step-by-step VPS example: exe.dev.

​
Install

​
Gateway service (systemd)

Install with one of: 
openclaw onboard --install-daemon
openclaw gateway install
openclaw configure   # select "Gateway service" when prompted
Repair or migrate an existing install: 
openclaw doctor
openclaw gateway install renders a systemd user unit by default. Full service guidance, including the system-level unit variant for shared or always-on hosts, lives in the Gateway runbook. Write a unit by hand only for a custom setup. Minimal user-unit example (~/.config/systemd/user/openclaw-gateway[-<profile>].service): 
[Unit]
Description=OpenClaw Gateway (profile: <profile>, v<version>)
After=network-online.target
Wants=network-online.target
StartLimitBurst=5
StartLimitIntervalSec=60

[Service]
ExecStart=/usr/local/bin/openclaw gateway --port 18789
Restart=always
RestartSec=5
RestartPreventExitStatus=78
TimeoutStopSec=30
TimeoutStartSec=30
SuccessExitStatus=0 143
OOMPolicy=continue
KillMode=control-group

[Install]
WantedBy=default.target
Enable it: 
systemctl --user enable --now openclaw-gateway[-<profile>].service

​
Memory pressure and OOM kills

On Linux, the kernel picks an OOM victim when a host, VM, or container cgroup runs out of memory. The Gateway is a poor victim because it owns long-lived sessions and channel connections, so OpenClaw biases transient child processes to be killed first when possible. For eligible Linux child spawns, OpenClaw wraps the command in a short /bin/sh shim that raises the child’s own oom_score_adj to 1000, then execs the real command. This is unprivileged: a process may always raise its own OOM score. Covered child process surfaces:
  • Supervisor-managed command children
  • PTY shell children
  • MCP stdio server children
  • OpenClaw-launched browser/Chrome processes (via the plugin SDK process runtime)
The wrapper is Linux-only and skipped when /bin/sh is unavailable, or when the child env sets OPENCLAW_CHILD_OOM_SCORE_ADJ to 0, false, no, or off. Verify a child process: 
cat /proc/<child-pid>/oom_score_adj
Expected value for covered children is 1000; the Gateway process itself keeps its normal score (usually 0). The systemd unit’s OOMPolicy=continue keeps the Gateway service alive when a transient child is selected by the OOM killer instead of marking the whole unit failed and restarting all channels; the failed child/session reports its own error. This does not replace normal memory tuning. If a VPS or container repeatedly kills children, raise the memory limit, reduce concurrency, or add stronger resource controls (systemd MemoryMax=, container memory limits).