Jarvis - Chief of AI Staff on NVIDIA GB10

Before running Jarvis, ensure:

1. Hardware: Dell Pro Max GB10 (NVIDIA DGX Spark) running Linux (Ubuntu-based) 2. OpenClaw: Installed and onboarded (openclaw --version returns a valid version) 3. LLM Backend: Either Ollama or LM Studio installed and running 4. Model: At least one model pulled (recommended: qwen3.5:35b-a3b or qwen3.5:27b) 5. Network: Tailscale VPN configured for remote access (recommended, not required) 6. Docker: Installed for SearXNG web search (docker --version)

Verify readiness:

# Check all prerequisites
echo "=== Jarvis Readiness Check ==="
echo "OS: $(uname -srm)"
echo "OpenClaw: $(openclaw --version 2>/dev/null || echo 'NOT INSTALLED')"
echo "Ollama: $(ollama --version 2>/dev/null || echo 'NOT INSTALLED')"
echo "Docker: $(docker --version 2>/dev/null || echo 'NOT INSTALLED')"
echo "GPU: $(nvidia-smi --query-gpu=name,memory.total --format=csv,noheader 2>/dev/null || echo 'NOT DETECTED')"
echo "Gateway: $(systemctl is-active openclaw 2>/dev/null || echo 'NOT RUNNING')"
echo "=== Check Complete ==="

Jarvis doesn't know who he is:

Memory is not persisting:

GPU overheating (>85°C):

WhatsApp not responding: