ClawMetry for OpenClaw

I have definitely run into situations where an agent is clearly doing something, but you have no idea what stage it is at or whether it is stuck, progressing, or just burning tokens. That uncertainty becomes frustrating pretty quickly.I would still want to understand how much control or intervention it allows beyond just observing what is happening.

Great execution on the real-time visibility problem. The sub-agent spawning scenario you describe is exactly where most logging tools fall apart.One gap I keep running into: tools that show what the agent did, but not what it was supposed to do at that step. The deviation between stated intent and actual execution is often where the real failure hides — and it's invisible in a standard trace.How are you thinking about capturing pre-action intent in ClawMetry, not just post-action results?

Hmm, nice visualization

Hi Vivek. I have my agent installed fully self-contained on a NAS, accessed via Slack from my Mac (see here for how I did it: https://moltbookagents.net/nas-setup-guide.html). Where do I install ClawMetry - on the NAS? How then do I access the interface? Thanks!

Brilliant, thank you for doing this

This is exactly the kind of tooling the AI agent ecosystem needs right now. I've been working on observability for MCP tool calls in production, and the biggest challenge is tracing what happens when an agent chains multiple actions — one wrong follow-up and the whole workflow breaks silently. The per-session cost breakdown is a smart move too. Most teams I've talked to have no idea what their agents actually cost until the invoice hits. Curious — are you planning to support tracing across agents that communicate with external APIs via webhooks or HTTP callbacks? That's where I see the next big observability gap opening up.

This work on OC running on Railway?

Introducing ClawMetry for OpenClaw 🚀

I built an real-time observability dashboard for OpenClaw AI agents. Here's why.

You send a message in your Telegram / WhatsApp to do some work & the agent says "spawning sub-agent" and you just... hope it works?

You shouldn't have to. So I built ClawMetry.

See what every sub-agent is doing. Right now.
What files it's reading, what commands it's running, what tools it's calling, what it's thinking. Summary, narrative, full logs. No more guessing.

System health at a glance.
Cron jobs, service uptime, disk usage, active sub-agents. All in one view.

Session history.
Every session logged with a timeline, tool calls, and cost. Go back and see exactly what happened.

Cost breakdown.
Per-session, per-model, per-tool. Know what you're spending before the invoice shows up.

It's open source, installs in 30 seconds:

pip install clawmetry

One line in your OpenClaw config, and you get a full dashboard.

If you're building with AI agents, you need visibility. Not more logs, actual observability.

clawmetry.com