ClawSpark: one-command private OpenClaw on NVIDIA DGX/RTX/Mac with zero telemetry

Developers and enterprises want to run OpenClaw agents on their own hardware without any data leaving the machine, but the setup requires manual LLM configuration, disabling telemetry flags, and wiring up local model servers. ClawSpark proved this demand exists but only supports NVIDIA DGX hardware. This tool auto-detects your GPU (NVIDIA, AMD, Apple Silicon), installs the optimal local LLM backend, configures OpenClaw for fully offline operation, and verifies zero network egress so you get private agents without the setup pain.