Documentation

1. Create your account

Sign up at clawpulse.org/signup. No credit card required -- all plans include a free trial.

2. Add your first instance

Go to Dashboard > Instances > Add Instance. Give it a name, host, and port. A unique token will be generated automatically.

3. Install the agent

Run this one-liner on the server where your OpenClaw instance runs. The agent installs as a systemd service and starts collecting telemetry automatically.

curl -sS https://www.clawpulse.org/agent.sh | sudo bash -s YOUR_TOKEN

4. Configure alerts

Go to Alerts > Destinations to add your Slack webhook, email, or Discord channel. Then create rules under Alerts > Rules to set thresholds.

System metrics

The agent collects: CPU usage, RAM usage, disk usage, swap, load averages (1m/5m/15m), open connections, running processes, and system uptime.

OpenClaw deep inspection

The agent discovers your OpenClaw process and inspects: PID, threads, file descriptors, sockets, LLM API connections, provider, model, tools, version, config files, and log files.

Live request metrics

Parsed from OpenClaw logs: requests/min, errors/min, average response time, active sessions, tokens in/out, and last error details.

API probing

The agent probes standard endpoints: /metrics, /health, /status, /api/health, /api/status, /api/metrics, /v1/health to detect API availability and latency.

Mission Control

Interactive real-time fleet visualization with animated agent nodes, health scoring (0-100), live event log, and task feed overlay. Nodes scale with request rate and show CPU arcs.

Monitoring

Two views: System (CPU/RAM/disk/swap gauges, 48h historical charts) and OpenClaw (process details, live metrics, request/error/token trend charts).

Task Tracker

Live task feed with intelligent grouping (message pairs, duplicate errors), detailed message cards (inbound, AI replies, errors, agent runs), browser notifications for critical errors.

Reports

Configurable period (7/14/30/90 days). Fleet KPIs, request/error trends, resource usage charts, token usage breakdown, task analytics by type/status, instance performance table. CSV export.

Supported channels

Slack (webhook), Email, WhatsApp (via Twilio), Discord (webhook). Each destination can be toggled on/off independently.

Alert rules

Create rules based on P95 latency thresholds. Assign severity (critical, high, medium, low) and notification channel. Rules can be muted/unmuted.

Telemetry endpoint

POST /api/dashboard/telemetry -- Receives system + OpenClaw telemetry from the agent. Authenticated via instance token.

Task feed endpoint

POST /api/dashboard/tasks -- Agent pushes individual task/request logs. GET returns paginated task history.

Health check

GET /api/health -- Returns { status: 'ok' }. Use this to monitor ClawPulse platform availability.