Which agent / tool talks to which brain — solid comet = routed through the ollama-proxy · dotted = straight to the GPU. Only the 4 OpenClaw agents go through the proxy. Live routing 2026-07-07.
| Primary | gemma4:26b-peepers — Hommer RX 7900 XTX |
| Fallbacks | qwen3-coder:30b-peepers (Hommer) → qwen3-coder:30b (1660 array) |
| Rule | Only agent allowed on Hommer; load → work → free VRAM |
| Model | qwen3.5:9b — RTX 3060 :11435 (ctx 32k, keep_alive 2h) |
| Fallback | qwen3-coder:30b — 1660 array |
| Rule | Never Hommer — dedicated RainbowAI box only |
| VRAM | 24 GB · fastest single GPU (91 tok/s qwen3-coder) |
| Runs | Peepers primary + fallback, aivid WAN 2.2 video |
| Note | Jess's workstation — freed after each job, never pinned |
| Array | 5×GTX 1660 Super + 1×1660 Ti · :11434 · qwen3-coder:30b (warm-kept) |
| RTX 3060 | 12 GB · :11435 · qwen3.5:9b + embeddings + SD images |
| Role | All non-Peepers inference + image gen |
| Routes | by model name → correct host/GPU |
| Injects | keep_alive per model; gemma4 /no_think |
| Guards | tool_call arg normalize + upstream-error passthrough → clean fallback (added 2026-07-06) |
| Model | Fable 5 / Opus 4.8 — Anthropic cloud |
| Role | Designs, builds & fixes the whole fleet; not a local model |
| Orchestrates | N8N, the bots, and this map itself |