ClawMetry x OpenInfer: run OpenClaw agents on any hardware and watch every turn

If you use ClawMetry to watch your OpenClaw agents, here is something new. We just shipped an integration with OpenInfer, which routes each inference workload to the most efficient hardware across its partner cloud.

The combination means you can run an agent across a mix of GPUs and CPUs and still see exactly what ran, where it ran, and what it cost, per session and per sub-agent. No code changes, still read-only.

Full write-up: https://clawmetry.com/blog/making-the-other-90-percent-observable

Try it: https://clawmetry.com/openinfer

Curious if anyone here is running on mixed hardware, would love to hear how it goes.