Hermes Agent

Multi-node Hermes Agent configuration, documentation, and shared resources.

Repository Structure

hermes-agent/
├── docs/
│   ├── README.md              # This file
│   ├── architecture.md        # Multi-agent architecture overview
│   ├── node-configs/
│   │   ├── strix.md           # strix-halo node configuration
│   │   └── bd790i.md          # bd790i node configuration
│   ├── migration/
│   │   └── bd790i-reconciliation.md  # bd790i migration from legacy setup
│   └── operations/
│       ├── docker-compose.md  # Docker compose patterns
│       └── llama-server.md    # llama-server configuration per node
├── templates/
│   ├── docker-compose.yml.template   # Base docker-compose template
│   └── config.yaml.template          # Base config template
└── scripts/
    └── deploy-node.sh              # Node deployment script

Nodes

strix-halo (192.168.50.16)

• Arch: x86_64
• Model: Qwen3.6-35B-A3B-MTP-UD-Q8_K_XL.gguf
• llama-server: server-vulkan on port 8000
• Status: ✅ Operational

bd790i (192.168.50.254)

• Arch: amd64
• Model: Qwen3.6-27B-MTP-Q4_K_M.gguf
• llama-server: server-cuda12 (llama-server-mtp) on port 8888
• Status: 🔄 Migration in progress

Quick Start

Each node runs hermes in a Docker container with:

• Shared config from hermes-shared repo (skills, scripts, plans, configs)
• Node-local data directory (sessions, state, cron, memories)
• Node-local .hermes directory (agent home, skills, auth)
• Local llama-server for inference

See docs/node-configs/ for per-node details.