Skip to main content
The Gateway is OpenClaw’s WebSocket server (channels, nodes, sessions, hooks). All subcommands below live under openclaw gateway ....

Bonjour discovery

Local mDNS + wide-area DNS-SD setup.

Discovery overview

How OpenClaw advertises and finds gateways.

Configuration

Top-level gateway config keys.

​
Run the Gateway

  • Refuses to start unless gateway.mode=local is set in ~/.openclaw/openclaw.json. Use --allow-unconfigured for ad-hoc/dev runs; it bypasses the guard without writing or repairing config.
  • Startup automatically applies deterministic, prompt-free legacy-key migrations to eligible invalid single-file configs, including in non-interactive service runs. It writes only after full validation, including plugins, and keeps the previous config in the .bak ring. Configs using $include, Nix-managed configs, and configs written by a newer version are excluded. See Legacy config key migrations.
  • If automatic migration cannot make the config valid, an interactive terminal can offer to run openclaw doctor --fix and retry startup once after consent. Non-interactive runs print the command instead. If the repaired config is still invalid, startup remains stopped.
  • openclaw onboard --mode local and openclaw setup write gateway.mode=local. If the config file exists but gateway.mode is missing, that is treated as damaged/clobbered config and the Gateway refuses to guess local for you — re-run onboarding, set the key manually, or pass --allow-unconfigured.
  • Binding beyond loopback without auth is blocked.
  • --bind values lan, tailnet, and custom resolve over IPv4-only paths today; IPv6-only bring-your-own-host setups need an IPv4 sidecar or proxy in front of the Gateway.
  • SIGUSR1 triggers an in-process restart when authorized. commands.restart (default: enabled) gates externally-sent SIGUSR1; set it to false to block manual OS-signal restarts. The agent-facing gateway tool is read-only; agents request restart through the openclaw delegation tool. Effective Full Access, including Default (Full Access), authorizes permitted delegated changes without an approval prompt; restricted runs require human approval. See Delegated setup and repair.
  • SIGINT/SIGTERM stop the process but do not restore custom terminal state — if you wrap the CLI in a TUI or raw-mode input, restore the terminal yourself before exit.

​
Options

​
number
WebSocket port (default from config/env; usually 18789).
​
string
Bind mode: loopback (default), lan, tailnet, auto, custom.
​
string
Shared token for connect.params.auth.token. Defaults to OPENCLAW_GATEWAY_TOKEN when set.
​
string
Auth mode: none, token, password, trusted-proxy.
​
string
Password for --auth password.
​
string
Read the Gateway password from a file.
​
string
Tailscale exposure: off, serve, funnel.
​
boolean
Start without enforcing gateway.mode=local. Ad-hoc/dev bootstrap only; does not persist or repair config.
​
boolean
Create a dev config + workspace if missing (skips BOOTSTRAP.md).
​
boolean
Allow the Gateway to auto-configure channels from ambient environment variables. By default, channels require an explicit channels.<id> config block.
​
boolean
Deprecated alias for --ambient-channels.
​
boolean
Reset dev config, credentials, sessions, and workspace. Requires --dev.
​
boolean
Kill any existing listener on the target port before starting. In a non-interactive shell, this refuses to kill a verified Gateway listener; use --dev or an isolated --profile with a free port instead.
​
boolean
Verbose logging to stdout/stderr.
​
boolean
Only show CLI backend logs in the console (also enables stdout/stderr).
​
string
default:"auto"
WebSocket log style: auto, full, compact.
​
boolean
Alias for --ws-log compact.
​
boolean
Log raw model stream events to JSONL.
​
string
Raw stream JSONL path.
--claude-cli-logs is a deprecated alias for --cli-backend-logs. For --bind custom, set gateway.customBindHost to an IPv4 address. Any address other than 127.0.0.1 or 0.0.0.0 also requires 127.0.0.1 on the same port for same-host clients; startup fails if either listener cannot bind. Wildcard 0.0.0.0 does not add a separate required alias. IPv6-only bring-your-own-host setups need an IPv4 sidecar or proxy in front of the Gateway.

​
Reveal the configured token

Run this on the Gateway host when a client needs the configured shared token:
The command resolves gateway.auth.token, OPENCLAW_GATEWAY_TOKEN, and configured SecretRefs, then prints only the token. It requires an interactive terminal and refuses redirected or piped output so the credential does not silently enter command logs. Treat the terminal output as a secret. If no persistent token is configured, run openclaw doctor --generate-gateway-token, restart the Gateway, and then rerun the command. Generic openclaw config get output remains redacted, including --json.

​
Restart the Gateway

--safe asks the running Gateway to preflight active work and schedule one coalesced restart after that work drains. The wait is bounded to 5 minutes; when the budget expires the restart is forced. --safe cannot combine with --force or --wait. --skip-deferral bypasses only the safe-restart active-work deferral gate. It can move the Gateway into shutdown even while active-work blockers are reported, but the close-stage pending-reply drain still applies before the process exits. It requires --safe — use it when a deferral is stuck on a runaway task and reply delivery can still be allowed to settle. --wait <duration> overrides the drain budget for a plain (non-safe) restart. Accepts bare milliseconds or unit suffixes ms, s, m, h, d (e.g. 30s, 5m, 1h30m); --wait 0 waits indefinitely. Not compatible with --force or --safe. --force skips the active-work drain and restarts immediately. Plain restart normally uses the service-manager restart path. On Windows, a plain restart launched from a Gateway service process, including an agent’s shell command, automatically uses the safe restart path. The running Gateway owns the deferred Scheduled Task handoff, so stopping its process tree cannot kill the caller before relaunch. This requires a reachable Gateway; the command acknowledges the restart request, not successor health. Use openclaw gateway status afterward to verify recovery. On macOS, when openclaw gateway restart, stop, install, or uninstall runs inside the managed LaunchAgent’s process tree, including an agent’s shell command, OpenClaw detects that from launchd’s service environment or, when a hand-written plist omits those variables, from process ancestry against the PID launchd reports for the job. Restart hands off to a detached helper so kickstart -k cannot kill the caller. Stop, install, and uninstall refuse and ask you to run the command from an external shell. External terminals without Gateway-service markers, externally supervised Gateways, node services, and non-Windows callers keep their existing routing. Explicit --force, --wait, --preserve-definition, or --skip-deferral also retain their existing behavior and validation; they do not implicitly enable --safe.
Inline --password can be exposed in local process listings. Prefer --password-file, env, or a SecretRef-backed gateway.auth.password.

​
Install identity

Service management (install, start, stop, restart, uninstall, Doctor service repair, and self-update service handling) belongs to the install that owns the host service. That is the canonical .openclaw directory under the OS account home, or the .openclaw-<profile> directory a named profile projects there. Named profiles use distinct native service identities. OPENCLAW_HOME, or an OPENCLAW_STATE_DIR or OPENCLAW_CONFIG_PATH that points elsewhere, is treated as isolated state and skipped. A relocated or copied state tree cannot adopt and rewrite the account’s host service. On macOS and Windows, native service-managed profile names must be lowercase. Runtime-only profiles may still use uppercase, but case-distinct names such as Main and main share paths on normal case-insensitive filesystems and cannot safely own separate native services. On macOS, the lowercase names gateway and node are also unavailable for native service management because their historical LaunchAgent labels collide with the default Gateway and node-host services. Named profiles must also use the native service identity derived from OPENCLAW_PROFILE. Unset OPENCLAW_LAUNCHD_LABEL, OPENCLAW_SYSTEMD_UNIT, or OPENCLAW_WINDOWS_TASK_NAME before service management; custom identities remain available for the default profile or runtime-only/external-supervisor setups. On Linux, openclaw gateway install --force refuses a sealed systemd service definition, or one whose write authority cannot be verified, before changing configuration, authentication tokens, or service files. The error keeps its SERVICE_DEFINITION_SEALED or SERVICE_DEFINITION_UNKNOWN prefix and adds a reason tag and next action, without printing private paths, config, environment values, or underlying inspection errors. For [unsafe-permissions], inspect the named artifact category locally. The service directory is ~/.config/systemd/user; on a fresh install, its nearest existing ancestor may be ~/.config. The service state directory belongs to the selected profile. Check directory metadata, not file contents:
Missing directories are normal on a fresh install. After confirming the affected path is yours and is not intentionally shared, remove group/other write access with chmod go-w <path> and retry the same command. Mode 0700 is appropriate for private directories. Do not recursively chmod, take ownership of system paths, or use sudo/--force to bypass the check. Foreign-owned files and sealed mounts require the deployment owner; inspection failures require restoring filesystem or native service-manager access first. Type-wide service.d defaults are inspected as shared read-only inputs and do not require write access. Root-owned selected units and unit-specific drop-ins remain protected.

​
External supervisors

Set OPENCLAW_SUPERVISOR_MODE=external only when another process manager owns the Gateway lifecycle. In this mode:
  • openclaw gateway restart preserves the existing safe, forced, and bounded-wait behavior while targeting the verified running Gateway instead of launchd, systemd, or Task Scheduler. Exact-lock restart delivery runs inside that Gateway, so a replacement CLI does not migrate shared state before the old process hands off.
  • Native service install, start, stop, and uninstall operations are refused with guidance to use the external supervisor.
  • OpenClaw self-update is refused so the supervisor can stop the Gateway, replace and finalize the runtime, and restart it safely.
  • A fresh-process restart writes a bounded SQLite handoff before clean exit. If persistence fails, the Gateway falls back to an in-process restart instead of exiting without a consumable handoff.
An external supervisor can also claim durable ownership of shared-state writes:
Before claiming, stop and verify every older Gateway, CLI, Doctor, updater, and native app process that can write the shared state database. Pre-contract processes do not understand the ownership row and cannot be retroactively fenced. Claim only after every remaining writer uses ownership-aware code and carries OPENCLAW_SUPERVISOR_MODE=external. The claim is idempotent for the same stable manager identifier and refuses a different manager. There is no automatic claim or unclaim path. Once claimed, unmarked writable shared-state opens fail before permissions, schema migration, additive repair, compaction, or other mutation. Read-only access remains available. This is protection against accidental unmarked same-user writers, not an authentication or lease protocol. For upgrades and rollbacks, have the supervisor create a consolidated WAL-consistent copied snapshot with no SQLite sidecars, then run the target release’s own openclaw database preflight <copied-state.sqlite> --json before activation. Numeric schema versions alone do not prove that a same-version additive shape is compatible. See Database schemas. OPENCLAW_SERVICE_REPAIR_POLICY=external remains a separate Doctor repair policy. It does not declare runtime ownership; supervisors that need both behaviors should set both variables. External supervisors can negotiate and consume restart handoffs through the hidden machine contract:
Protocol version 1 supports the consume operation. Consumption validates the expected PID and bounded handoff fields inside one immediate SQLite transaction. An accepted handoff is deleted before success is returned, so concurrent or replayed consumers cannot both accept it. A PID mismatch is retained for the matching owner; missing, expired, and invalid rows do not authorize a restart. Valid machine requests return JSON with exit code 0, including non-restart results. Invalid arguments return reason: "invalid-expected-pid" with exit code 2; state-store failures return reason: "store-unavailable" with exit code 1. Supervisors should probe capabilities on the exact runtime or launcher they will use rather than infer support from an OpenClaw version string or read the private SQLite schema directly. External supervisor implementations should also apply these acceptance rules:
  • Bound capability probes with a timeout that accounts for full CLI cold-start latency on the deployed runtime and storage, rather than assuming warm-start timing.
  • If capability negotiation or handoff consumption refuses replacement, exit promptly with a nonzero status so the process manager’s recovery policy can run. Do not remain alive without a Gateway child or listener.
  • Treat supervisor process liveness as distinct from replacement startup and channel readiness. Report success only after the new Gateway owns its listener and /startupz returns status: "started"; monitor /readyz separately for configured-channel health, while /healthz proves liveness only.

​
Gateway profiling

  • OPENCLAW_GATEWAY_STARTUP_TRACE=1 logs phase timings during startup, including per-phase eventLoopMax delay and plugin lookup-table timings (installed-index, manifest registry, startup planning, owner-map work).
  • OPENCLAW_GATEWAY_RESTART_TRACE=1 logs restart trace: lines for restart signal handling, active-work drain, shutdown phases, next start, ready timing, and memory metrics. Ordinary stops also start a fresh trace with stop.signal.received and stop.drain timing. Named shutdown steps and coarse close phases emit .begin before waiting, then a duration when they settle; an unmatched begin identifies an entered phase that has not settled. These phases do not time every nested cleanup operation individually.
  • OPENCLAW_DIAGNOSTICS=timeline with OPENCLAW_DIAGNOSTICS_TIMELINE_PATH=<path> writes a best-effort JSONL startup diagnostics timeline for external QA harnesses (equivalent to config diagnostics.flags: ["timeline"]; the path is still env-only). Add OPENCLAW_DIAGNOSTICS_EVENT_LOOP=1 to include event-loop samples.
  • pnpm build then pnpm test:startup:gateway -- --runs 5 --warmup 1 benchmarks Gateway startup against the built CLI entry: first process output, /healthz, /readyz, startup trace timings, event-loop delay, and plugin lookup-table timing.
  • pnpm build then pnpm test:restart:gateway -- --case skipChannels --runs 1 --restarts 5 benchmarks in-process restart on macOS or Linux (not supported on Windows; restart requires SIGUSR1). Uses SIGUSR1, enables both traces in the child process, and records next /healthz, next /readyz, downtime, ready timing, CPU, RSS, and restart trace metrics.
  • /healthz is liveness; /readyz is usable readiness. Treat trace lines and benchmark output as owner-attribution signal, not a complete performance conclusion from one span or sample.
Without tracing, stops and restarts report nonzero active-work category counts at the first drain snapshot and at most once every 30 seconds while still pending. These reports omit task identities and request origins; categories can overlap. An ordinary stop logs active-work drain settled; beginning server close before teardown, including after a drain timeout or failure. Diagnostics do not change drain budgets or the service manager’s stop deadline. A client disconnect leaves interactive setup available for reconnect. A Gateway stop or restart closes setup prompts before draining work. Settings writes already in progress may finish, but setup will not wait for another answer during shutdown. After the Gateway starts again, reopen setup and check the saved settings.

​
Query a running Gateway

All query commands use WebSocket RPC.
  • Default: human-readable (colored in TTY).
  • --json: machine-readable JSON (no styling/spinner).
  • --no-color (or NO_COLOR=1): disable ANSI while keeping human layout.
When you set --url, the CLI does not fall back to config or environment credentials. Pass --token or --password explicitly. Missing explicit credentials is an error.

​
gateway health

/healthz is a liveness probe: it returns as soon as the server can answer HTTP. /readyz is stricter and stays red while startup plugin sidecars, channels, or configured hooks are still settling. Local or authenticated detailed /readyz responses include an eventLoop diagnostic block (delay, utilization, CPU-core ratio, degraded flag).
​
number
Target a local loopback Gateway on this port. Overrides OPENCLAW_GATEWAY_URL and OPENCLAW_GATEWAY_PORT for this call.

​
gateway usage-cost

Fetch usage-cost summaries from session logs.
​
number
default:"30"
Number of days to include.
​
string
Scope the summary to one configured agent id.
​
boolean
Aggregate across all configured agents. Cannot combine with --agent.

​
gateway stability

Fetch the recent diagnostic stability recorder from a running Gateway.
​
number
default:"25"
Maximum recent events to include (max 1000).
​
string
Filter by diagnostic event type, e.g. payload.large or diagnostic.memory.pressure.
​
number
Include only events after a diagnostic sequence number.
​
string
Read a persisted stability bundle instead of calling the running Gateway. --bundle latest (or bare --bundle) picks the newest bundle under the state directory; you can also pass a bundle JSON path directly.
​
boolean
Write a shareable support diagnostics zip instead of printing stability details.
​
string
Output path for --export.
  • Records keep operational metadata: event names, counts, byte sizes, memory readings, queue/session state, approval ids, channel/plugin names, and redacted session summaries. They exclude chat text, webhook bodies, tool outputs, raw request/response bodies, tokens, cookies, secret values, hostnames, and raw session ids. Set diagnostics.enabled: false to disable the recorder entirely.
  • Fatal Gateway exits, shutdown timeouts, and restart startup failures write the same diagnostic snapshot to ~/.openclaw/logs/stability/openclaw-stability-*.json when the recorder has events. Inspect the newest bundle with openclaw gateway stability --bundle latest; --limit, --type, and --since-seq apply to bundle output too.

​
gateway diagnostics export

Write a local diagnostics zip designed for bug reports. For the privacy model and bundle contents, see Diagnostics Export.
​
string
Output zip path. Defaults to a support export under the state directory.
​
number
default:"5000"
Maximum sanitized log lines to include.
​
number
default:"1000000"
Maximum log bytes to inspect.
​
string
Gateway WebSocket URL for the health snapshot.
​
string
Gateway token for the health snapshot.
​
string
Gateway password for the health snapshot.
​
number
default:"3000"
Status/health snapshot timeout.
​
boolean
Skip persisted stability bundle lookup.
​
boolean
Print the written path, size, and manifest as JSON.
The export bundles: manifest.json (file inventory), summary.md (Markdown summary), diagnostics.json (top-level config/logs/discovery/stability/status/health summary), config/sanitized.json, status/gateway-status.json, health/gateway-health.json, logs/openclaw-sanitized.jsonl, and stability/latest.json when a bundle exists. It is designed to be shared. It keeps operational details useful for debugging — safe log fields, subsystem names, status codes, durations, configured modes, ports, plugin/provider ids, non-secret feature settings, and redacted operational log messages — and omits or redacts chat text, webhook bodies, tool outputs, credentials, cookies, account/message identifiers, prompt/instruction text, hostnames, and secret values. When a log message looks like user/chat/tool payload text (e.g. “user said”, “chat text”, “tool output”, “webhook body”), the export keeps only the fact that a message was omitted plus its byte count.

​
gateway status

Shows the Gateway service (launchd/systemd/schtasks) plus an optional connectivity/auth probe.
​
string
Probe this explicit WebSocket URL instead of the service-derived target. Cannot combine with --port.
​
number
Select a local Gateway port using the invoking CLI config for auth and TLS. Accepts gateway --port 19001 status and gateway status --port 19001; an explicit status port wins. Native service details remain visible as diagnostics but do not select the probe target.
​
string
Token auth for the probe.
​
string
Password auth for the probe.
​
number
default:"10000"
Probe timeout.
​
boolean
Skip the connectivity probe (service-only view).
​
boolean
Scan system-level services too.
​
boolean
Upgrade the connectivity probe to a read probe and exit non-zero if it fails. Cannot combine with --no-probe.
  • Stays available for diagnostics even when the local CLI config is missing or invalid.
  • Default output proves service state, WebSocket connect, and the auth capability visible at handshake time — not read/write/admin operations.
  • Probes are non-mutating for first-time device auth: they reuse an existing cached device token when one exists, but never create a new CLI device identity or read-only pairing record just to check status.
  • Resolves configured auth SecretRefs for probe auth when possible. If a required SecretRef is unresolved, --json reports rpc.authWarning when probe connectivity/auth fails; pass --token/--password explicitly or fix the secret source. Unresolved-auth warnings are suppressed once the probe succeeds.
  • JSON output includes gateway.version when the running Gateway reports it; --require-rpc can fall back to the status.runtimeVersion RPC payload if the handshake probe cannot supply version metadata.
  • Use --require-rpc in scripts/automation when a listening service is not enough and you need read-scope RPC to be healthy too.
  • --deep scans for extra launchd/systemd/schtasks installs; when multiple gateway-like services are found, human output prints cleanup hints (usually run one gateway per machine) and reports a recent supervisor restart handoff when relevant.
  • --deep confirms exact npm targets before suggesting repairs for official-plugin version drift. Unpublished versions or registry failures are reported without an update command; retry deep status after registry access or the release cohort is restored. Ordinary status and readiness checks do not query npm for drift repairs.
  • --deep also runs config validation in plugin-aware mode (pluginValidation: "full") and surfaces plugin manifest warnings (e.g. missing channel config metadata). Default gateway status keeps the fast read-only path that skips plugin validation.
  • On Linux, status reports the effective service currently loaded by systemd, including loaded drop-ins. If the unit or a drop-in changed on disk, Systemd reload: pending means you must run systemctl --user daemon-reload (or sudo systemctl daemon-reload for a system service) before those changes take effect.
  • Human output includes the resolved file log path plus CLI-vs-service config paths/validity to help diagnose profile or state-dir drift.
  • Install and reinstall guidance follows the invoking shell’s installation rules, not the stored service environment or probe target. Nix mode, external supervision, noncanonical installation identity, and Linux sudo/user-manager mismatches show the install refusal instead of an unusable command. A diagnostic-only target is not itself a refusal. Nix mode blocks installation, not starting an existing service.
  • Human output includes Gateway heap: with configured service heap controls and a separate install-time recommendation based on memory visible to the CLI. JSON output exposes the same report as service.gatewayHeap. Neither is a measurement of the running Gateway’s V8 heap ceiling; use runtime memory diagnostics for that.
  • Service auth drift checks read both Environment= and EnvironmentFile= from the unit (including %h, quoted paths, multiple files, and optional - files).
  • Resolves gateway.auth.token SecretRefs using merged runtime env (service command env first, then process env fallback).
  • Token-drift checks skip config token resolution when token auth is not effectively active (gateway.auth.mode explicitly password/none/trusted-proxy, or mode unset where password can win and no token candidate can win).

​
gateway probe

The “debug everything” command. It always probes:
  • your configured remote gateway (if set), and
  • localhost (loopback), even if remote is configured.
Passing --url adds that explicit target ahead of both. Human output labels targets URL (explicit), Remote (configured) / Remote (configured, inactive), and Local loopback.
If multiple probe targets are reachable, all are printed. An SSH tunnel, TLS/proxy URL, and configured remote URL can point at the same gateway even with different transport ports; multiple_gateways is reserved for distinct or identity-ambiguous reachable gateways. Running multiple gateways is supported for isolated profiles (e.g. a rescue bot), but most installs run a single gateway.
​
number
Use this port for the local loopback probe target and SSH tunnel remote port. Without --url, this selects only the local loopback target instead of configured gateway environment URL, environment port, or remote targets.
  • Reachable: yes means at least one target accepted a WebSocket connect.
  • Capability: read-only|write-capable|admin-capable|pairing-pending|connect-only reports what the probe could prove about auth, separate from reachability.
  • Read probe: ok means read-scope detail RPC calls (health/status/system-presence/config.get) also succeeded.
  • Read probe: limited - missing scope: operator.read means connect succeeded but read-scope RPC is limited. Reported as degraded reachability, not full failure.
  • Read probe: failed after Connect: ok means the WebSocket connected but follow-up read diagnostics timed out or failed — also degraded, not unreachable.
  • Like gateway status, probe reuses existing cached device auth but does not create first-time device identity or pairing state.
  • Exit code is non-zero only when no probed target is reachable.
Top level:
  • ok: at least one target is reachable.
  • degraded: at least one target accepted a connection but did not complete full detail RPC diagnostics.
  • capability: best capability seen across reachable targets (read_only, write_capable, admin_capable, pairing_pending, connected_no_operator_scope, or unknown).
  • primaryTargetId: best target to treat as the active winner, in order: explicit URL, SSH tunnel, configured remote, local loopback.
  • warnings[]: best-effort warning records with code, message, optional targetIds.
  • network: local loopback/tailnet URL hints derived from current config and host networking.
  • discovery.timeoutMs / discovery.count: the actual discovery budget/result count used for this probe pass.
Per target (targets[].connect): ok (reachability + degraded classification), rpcOk (full detail RPC success), scopeLimited (detail RPC failed on missing operator scope).Per target (targets[].auth): role and scopes reported in hello-ok when available, plus the surfaced capability classification.
  • ssh_tunnel_failed: SSH tunnel setup failed; the command fell back to direct probes.
  • multiple_gateways: distinct gateway identities were reachable, or OpenClaw could not prove reachable targets are the same gateway. An SSH tunnel, proxy URL, or configured remote URL to the same gateway does not trigger this.
  • auth_secretref_unresolved: a configured auth SecretRef could not be resolved for a failed target.
  • probe_scope_limited: WebSocket connect succeeded, but the read probe was limited by missing operator.read.
  • local_tls_runtime_unavailable: local Gateway TLS is enabled but OpenClaw could not load the local certificate fingerprint.

​
Remote over SSH (Mac app parity)

The macOS app “Remote over SSH” mode uses a local port-forward so a loopback-only remote gateway becomes reachable at ws://127.0.0.1:<port>. CLI equivalent:
​
string
user@host or user@host:port (port defaults to 22).
OpenClaw launches only an SSH client found in OS-managed system directories. On native Windows, install the OpenSSH Client optional feature; Windows places it under %SystemRoot%\System32\OpenSSH.
​
string
Identity file.
​
boolean
Pick the first discovered gateway host as SSH target from the resolved discovery endpoint (local. plus the configured wide-area domain, if any). TXT-only hints are ignored.
Config defaults (optional): gateway.remote.sshTarget, gateway.remote.sshIdentity.

​
gateway call <method>

Low-level RPC helper.
​
string
default:"{}"
JSON object string for params.
​
string
Gateway WebSocket URL.
​
number
Target a local loopback Gateway on this port. Overrides OPENCLAW_GATEWAY_URL and OPENCLAW_GATEWAY_PORT for this call. Cannot combine with --url.
​
string
Gateway token.
​
string
Gateway password.
​
number
default:"10000"
Timeout budget.
​
boolean
Mainly for agent-style RPCs that stream intermediate events before a final payload.
​
boolean
Machine-readable JSON output.
openclaw.setup.detect uses a 40-second default so the Gateway can finish its bounded AI-access scan. An explicit --timeout still takes precedence.
--params must be valid JSON, and each method validates its own param shape (extra/misnamed fields are rejected). Use --port for a custom-port local Gateway; explicit --url targets still require explicit credentials.

​
gateway suspend

Prepare an idle Gateway for a cooperative host freeze or snapshot. Without --wait, active work returns a nonzero exit with blocker details. With --wait, the CLI retries until the bounded deadline using one stable request ID. The value must be a non-negative number of seconds; an empty value is rejected. Use --wait 0 for a single attempt without polling.
The ready output includes the suspension ID, lease expiry, and the matching resume command. Common RPC options such as --url, --token, --password, --timeout, --json, and --port are supported.

​
gateway resume <suspensionId>

Release a prepared suspension after thaw or when the host operation is abandoned.
An already expired or resumed lease is a successful no-op. A different active suspension ID is rejected.

​
Manage the Gateway service

​
Recover an unreadable native service definition

If installation or a managed update reports SERVICE_DEFINITION_UNKNOWN, first restore access to the service files and native service manager. --force does not bypass unknown service facts. Inspect the selected service with openclaw gateway status --deep from the account and profile that own it. For a malformed definition or unsupported environment syntax, privately back up the service files and any values stored only in its environment. Correct unresolved or unsupported values before reinstalling; OpenClaw cannot infer their intended values. Then, from an external shell using the same account and profile:
Uninstall removes the native registration and launcher, preserving configuration, plugin installations, session state, and workspaces. Reinstallation rebuilds the service environment from the current configuration and installation inputs; service-only values must be supplied again. If native service status is itself unavailable, uninstall also refuses: restore native manager access first instead of deleting state or bypassing inspection.

​
Lifecycle requests from Gateway chat

Gateway-hosted OpenClaw chat controls the exact Gateway serving that session. An approved start request reports Gateway already running without discovering or starting another service. Restart keeps the safe local restart behavior. An approved stop reports Scheduled Gateway stop after the host has prepared the stop for its exact instance. This acknowledges scheduling, not completed termination. An exclusive foreground host drains work, finishes teardown, and exits successfully without discovering or changing an installed service. A host managed by launchd or systemd verifies native ownership and prepares an executor, then drains work and finishes teardown before asking the native manager to stop the service. The requesting operation can finish its audit, history, and response submission during that drain; this does not guarantee that the client receives the response before disconnecting. After the normal grace period, stop cancels the remaining runs owned by that Gateway and waits for their commands and cleanup to settle. Ordinary stop does not schedule restart recovery. A required cleanup failure produces a nonzero exit and prevents an in-process replacement, including when startup failed before the Gateway became ready. Ownership or preparation failures leave the Gateway serving and return an error. Linux uses an independent transient control scope, in the owning systemd manager, so the stop command survives service cgroup termination. On macOS, hosted stop requests ordinary launchctl bootout without changing persistent enablement. If the native manager sends SIGTERM during the final stop handoff, the host finishes its graceful exit after joining cleanup, including the owned stop client. On Windows, a run loop that exclusively owns the Gateway process also uses graceful process exit under Task Scheduler. It does not select or stop a task by name. The generated task supervisor waits for the child process tree to exit and propagates the child exit result through the launcher. Its RestartOnFailure policy does not restart a successful task exit. Custom wrappers can have different exit or restart behavior; check their policy separately. This stop path does not change the task definition or its restart policy. Externally supervised Gateways direct stop requests to their supervisor. If systemd definitively refuses a stop after teardown and the same native instance remains active with no pending job, the host logs the failed stop and starts a fresh Gateway generation in the same process. An uncertain native result is recorded as a failed shutdown, without claiming success or starting an in-process replacement. After an unexpected disconnect, check openclaw gateway status and the native service logs from an external shell before retrying. Standalone CLI lifecycle commands retain their service-management behavior.

​
Install with a wrapper

Use --wrapper when the managed service must start through another executable, for example a secrets manager shim or a run-as helper. The wrapper receives the normal Gateway args and is responsible for eventually exec’ing openclaw or Node with those args.
You can also set the wrapper through the environment. gateway install validates that the path is an executable file, writes the wrapper into the service ProgramArguments, and persists OPENCLAW_WRAPPER in the service environment for later forced reinstalls, updates, and doctor repairs.
To remove a persisted wrapper, clear OPENCLAW_WRAPPER while reinstalling:
  • gateway status: --url, --port, --token, --password, --timeout, --no-probe, --require-rpc, --deep, --json
  • gateway install: --port, --runtime <node|bun> (default: node), --token, --wrapper <path>, --force, --json
  • gateway restart: --safe, --skip-deferral, --force, --wait <duration>, --preserve-definition, --json
  • gateway uninstall|start: --json
  • gateway stop: --disable, --force, --json
  • Node is the primary, default, and recommended managed Gateway runtime.
  • Bun 1.4+ with WAL-reset-safe node:sqlite is available as an explicit opt-in with gateway install --runtime bun.
  • gateway start is idempotent: when the managed service is already running, it reports the running process and leaves it untouched. A loaded but stopped service is started as before.
  • If no managed service is installed, gateway start prints install hints and exits nonzero. gateway restart can first recover an installed-but-unloaded LaunchAgent or a verified unmanaged Gateway; if neither a managed service nor recovery handles the action, it prints the same hints and exits nonzero. Stopping an absent service remains a successful no-op.
  • If gateway start or gateway restart needs to repair a stale service definition, the command refuses when the invoking shell resolves a different state directory, config path, or port than the installed service. Match or unset the conflicting environment overrides, or use openclaw gateway install --force to retarget the service intentionally.
  • On Linux, gateway start and gateway restart also refuse ineffective repairs when an operator-owned systemd drop-in overrides the command or working directory. Inspect the effective unit with systemctl --user cat <unit>.service, then update or remove that drop-in. gateway install --force rewrites only the managed base unit and warns if the override remains; Environment= drop-ins remain supported.
  • gateway restart --preserve-definition restarts only an inspectable native service, skips automatic definition repair, and checks health at the installed launcher’s port. It does not recover an unmanaged listener and cannot be combined with --safe or external supervision. On macOS it can bootstrap an unloaded readable plist without rewriting the plist, environment, wrapper, or permissions; denied native activation fails without file repair. On Windows it also retains existing Startup entries. The legacy daemon restart command accepts the same option. Older CLIs reject the option before running restart or repair.
  • During writable Linux service installs or refreshes, keep the unit and state directories stationary and avoid concurrent manual edits. OpenClaw serializes its own writers and aborts on detected changes, but cannot coordinate arbitrary filesystem edits. Moving or replacing a parent directory mid-publication can leave a temporary file inside the moved directory; inspect it before retrying.
  • Use gateway restart to restart a managed service. Do not chain gateway stop and gateway start as a restart substitute.
  • In a non-interactive shell, gateway stop requires --force. Interactive terminals keep the existing prompt-free behavior. For automation and tests, prefer gateway run --dev or an isolated --profile with a free port.
  • On macOS, gateway stop uses launchctl bootout by default, which removes the LaunchAgent from the current boot session without persisting a disable — KeepAlive auto-recovery stays active for future crashes and gateway start re-enables cleanly without a manual launchctl enable. Pass --disable to persistently suppress KeepAlive and RunAtLoad so the gateway does not respawn until the next explicit gateway start; use this when a manual stop should survive reboots.
  • Gateway lifecycle mutations append best-effort key-value audit records to <state-dir>/logs/gateway-restart.log, including CLI start, stop, and restart operations, safe restart requests, supervisor restarts, and detached handoffs.
  • Lifecycle commands accept --json for scripting.
  • For a managed Node Gateway without an existing heap setting, gateway install places --max-old-space-size in Node’s launch arguments, before the entry script. It explicitly clears NODE_OPTIONS in the service environment so ambient service-manager preload/debug flags cannot leak into the Gateway. Plain spawned Node processes do not inherit the new automatic budget through NODE_OPTIONS; Node’s fork and Worker inheritance rules are unchanged.
  • Capacity is the smaller of valid physical RAM and a valid constraint reported by Node, never fluctuating free RAM. With no usable capacity reading, Node keeps its native default. The installer targets 50% of capacity, with a nominal 2048 MiB floor and a cap of the greater of 8192 MiB or 25% of capacity. A final 75% capacity cap reserves native-memory headroom and can put small-host budgets below the nominal floor.
  • Examples: 32 GiB capacity selects 8 GiB old space; 64 GiB selects 16 GiB; 128 GiB selects 32 GiB. Old space is only part of V8’s total heap, and neither is a limit on total process memory (RSS). Raising the ceiling does not preallocate that memory.
  • Existing managed service heap controls are preserved across forced reinstalls and doctor repairs, including absolute old-space, percentage old-space, and total-heap flags. Only heap flags survive managed NODE_OPTIONS sanitization; arbitrary preload/debug flags do not. Put intentional preload/debug settings in an operator-owned systemd Environment= drop-in, or set them inside an installed wrapper before it launches Node. Do not edit the generated service environment for those settings. Existing stored numbers are preserved even when they resemble an older automatic default or exceed the new recommendation.
  • When an operator-owned service override controls NODE_OPTIONS (including an empty value or reset), regeneration does not add a new automatic heap argument. Operator values and drop-in files stay separate from the managed base. Existing managed argv controls remain: Node’s argv wins over NODE_OPTIONS for the same option, and percentage old-space sizing takes precedence over absolute old-space sizing. Inspect both surfaces before changing a cap.
  • Ambient installer NODE_OPTIONS and the installer’s own Node arguments are not saved as Gateway heap settings. The budget is chosen at installation and takes effect when the service process starts; it is not recalculated while the Gateway runs. Upgrading OpenClaw alone does not resize a running Gateway, and foreground launches do not replace themselves to apply this policy.
  • The installer’s memory constraints can differ from the future service’s constraints. Node/libuv reporting is platform-dependent and does not guarantee detection of every ancestor cgroup limit; inspect the actual service or container limits before increasing a budget.
  • This policy applies to managed Node Gateway launches, not foreground gateway run, custom supervisors, Docker runtime commands, Bun, or node-host services. Those retain their own runtime configuration. See memory troubleshooting for explicit native Node settings.
  • When token auth requires a token and gateway.auth.token is SecretRef-managed, gateway install validates that the SecretRef is resolvable but does not persist the resolved token into service environment metadata.
  • If token auth requires a token and the configured token SecretRef is unresolved, install fails closed instead of persisting fallback plaintext.
  • For password auth on gateway run, prefer OPENCLAW_GATEWAY_PASSWORD, --password-file, or a SecretRef-backed gateway.auth.password over inline --password.
  • In inferred auth mode, shell-only OPENCLAW_GATEWAY_PASSWORD does not relax install token requirements; use durable config (gateway.auth.password or config env) when installing a managed service.
  • If both gateway.auth.token and gateway.auth.password are configured and gateway.auth.mode is unset, install is blocked until mode is set explicitly.

​
Discover gateways (Bonjour)

gateway discover scans for Gateway beacons (_openclaw-gw._tcp).
  • Multicast DNS-SD: local.
  • Unicast DNS-SD (wide-area Bonjour): choose a domain (example: openclaw.internal.) and set up split DNS + a DNS server; see Bonjour.
Only gateways with Bonjour discovery enabled (default) advertise the beacon. TXT hints on every beacon: role (gateway role hint), transport (transport hint, e.g. gateway), gatewayPort (WebSocket port, usually 18789), tailnetDns (MagicDNS hostname, when available), gatewayTls / gatewayTlsSha256 (TLS enabled + cert fingerprint). sshPort and cliPath are published only in full discovery mode (discovery.mdns.mode: "full"; default is "minimal", which omits them — clients then default SSH targets to port 22).

​
gateway discover

​
number
default:"2000"
Per-command timeout (browse/resolve).
​
boolean
Machine-readable output (also disables styling/spinner).
Examples:
  • Scans local. plus the configured wide-area domain when one is enabled.
  • wsUrl in JSON output is derived from the resolved service endpoint, not from TXT-only hints such as lanHost or tailnetDns.
  • discovery.mdns.mode controls sshPort/cliPath publication on both local. mDNS and wide-area DNS-SD (see above).