Self-healing monitor for OpenClaw. Watches your gateway, crons, and sessions. Alerts you on Telegram. Auto-fixes what it can.
ClawDoctor Status ───────────────── Daemon: ● running Telegram: ✓ enabled Running quick check... ✓ [GatewayWatcher] Gateway running ! [CronWatcher] Cron 'writer-content-drafts' last delivery failed ✓ [CronWatcher] 12 cron job(s) healthy ✓ [SessionWatcher] 23 recent session(s) healthy ✓ [AuthWatcher] No auth failures detected ✓ [CostWatcher] Cost baseline normal
Real output from our own OpenClaw setup. It caught 4 cron delivery failures on first run.
Five watchers running on intervals, reading directly from OpenClaw's state files.
Checks if your OpenClaw gateway process is running. Auto-restarts if it goes down.
Did your cron run? Did it succeed? Was the output actually delivered?
Catches aborted sessions, stuck loops, and sessions that ran way too long.
Detects 401/403 patterns and expired tokens before they cascade.
Flags when a session burns 3x more than your baseline. Catches runaway spend.
Not just alerts. ClawDoctor takes action.
Auto-restarts via openclaw gateway restart, verifies it's back, sends you a Telegram alert. You never knew it happened.
Logs the failure, includes the exact rerun command in the alert. No more SSHing in to figure out which cron broke or when.
OAuth token expired? ClawDoctor refreshes it. Config drift? Snapshots and repairs. Risky fix? Asks you on Telegram first.
Works on Linux. macOS coming soon. Requires Node.js 18+.
Intro pricing for first 100 customers. Locked in for 12 months.
Open source, forever
Intro price $15
Intro price $39
“Sentry monitors your cron jobs. Langfuse monitors your LLM calls. Nobody monitors whether your OpenClaw setup is actually working.”
Built by people who run 20+ OpenClaw agents in production.