A practical Hermes setup with manager and receptionist roles

The local uses openclaw together with Hermes. Each agent runs on a tiny PC with almost no computing power of its own. The thinking step runs on a separate RTX 3090 machine using GLM 4.7 Flash.

Heavier work runs on on a Jetson Thor AGX with 128 GB of . Coding uses Qwen3 Coder Next, while specialist jobs such as research and code review by a different model family use separate models. One manages the openclaw agents.

Another acts like a secretary or receptionist for the messaging contact point. The open question is whether better model choices or role assignments would make this setup more effective.

Key points

  • Use one as the manager for openclaw agents.
  • Use another as the secretary or receptionist for incoming messages.
  • Run lightweight thinking on a separate RTX 3090 machine with GLM 4.7 Flash.
  • Run heavier model work on a Jetson Thor AGX with 128 GB of .
  • Split work by role, such as Qwen3 Coder Next for coding and separate models for research or code review.
Read original