Observability and governance for OpenClaw

See every OpenClaw session, channel message, cron run, and token as it happens. ClawMetry detects OpenClaw automatically and renders it all in one dashboard. Free forever for OpenClaw. No SDK. No code change.

What ClawMetry sees on OpenClaw

OpenClaw records a transcript for every session and every chat channel, and streams live events over the gateway WebSocket (JSON-RPC). ClawMetry ingests all three sources and surfaces:

• Sessions: every agent session with model, started at, last activity, message count, and per-session token totals.
• Chat channels: conversations across 21 adapters (Telegram, Signal, WhatsApp, Discord, Slack, IRC, iMessage, WebChat, and more), with per-channel volume and cost.
• Live activity: every tool call, message, and thinking block as it lands, streamed from the gateway.
• Crons: full cron management (create, edit, pause, delete via gateway RPC) plus a run log with duration, outcome, and cost per run.
• Sub-agents: spawned sub-agents shown as a tree with each child's own tokens and tools rolled up under the parent.
• Cost per session, per model, per hour: cache-aware token accounting, comparable side by side with the other 19 runtimes ClawMetry observes.
• Approval + kill switch: token-maxing signals (message repeat, no-progress heuristic, cost-per-hour threshold) flag a session so you can pause or kill it before the next call.

Enable it in about two minutes

1. Install ClawMetry

Pick the line that matches your OS. Windows users: paste the install.cmd line, not the shell one.

# macOS or Linux
curl -fsSL https://clawmetry.com/install.sh | bash

# Windows (cmd.exe or PowerShell)
curl -fsSL https://clawmetry.com/install.cmd -o install.cmd && install.cmd

Already have Python? pip works too; onboard is the interactive first-run wizard.

# Any OS with Python 3.9+
pip install clawmetry && clawmetry onboard

Or skip the terminal entirely and grab the desktop app: macOS (.dmg) · Windows (.exe) · Linux (.AppImage). It installs and updates ClawMetry for you.

2. Run the daemon

clawmetry

The daemon auto-detects OpenClaw and the gateway, and starts reading sessions, channels, and logs immediately. No env var, no config, no restart of OpenClaw.

3. Open the dashboard

Open http://localhost:8900. Your OpenClaw node is already there.

Start free →

Free forever for OpenClaw

OpenClaw observability is free on every ClawMetry plan, on your own machine, with no time limit. Paid plans (Cloud Pro and Self-Hosted Pro) add the other 19 runtimes, the cloud dashboard, alerts, and multi-node fleet views. See pricing for the full split.

Governance, not just graphs

Watching an autonomous agent is only half the job. ClawMetry adds the guardrails a long-running OpenClaw node needs: an approval queue for risky tool calls, token-maxing detection for runaway loops, budget alerts before the invoice surprises you, and a kill switch that stops a session before its next call. All of it runtime-aware, so an OpenClaw cron gone rogue is flagged the same way a stuck sub-agent is.

What each tab shows for OpenClaw

Tab	What it shows	Scope
Sessions	Every OpenClaw session, sorted by last activity, with model and token totals	All sessions, or one
Channels	Conversations per chat adapter with volume and cost	21 adapters
Brain	Live entries as they land: messages, tool calls, sub-agent spawns, thinking	Whole node, or one session
Cost	Per-session, per-model, per-hour token spend; cache-read vs uncached-input split	Runtime = openclaw, or all
Crons	Cron CRUD plus run log with duration, outcome, and cost	Per node
Approvals	Pause + kill switch for token-maxing sessions and risky tool calls	Runtime-filtered

Starred by the people building AI

Engineers and founders from OpenAI, Google, PostHog and more have starred ClawMetry on GitHub. One of them helped build it.

All 20 runtimes ClawMetry observes

OpenClaw is one of 20. Free on every plan: OpenClaw, NVIDIA NemoClaw. Paid on Cloud Pro or Self-Hosted Pro: the rest.