npx -y @tencent-weixin/openclaw-weixin-cli@latest install
启动模型
vllm serve /opt/openclaw_master/models/Qwen3-4B-Instruct --trust-remote-code --max-model-len 16384 --gpu-memory-utilization 0.90 --dtype auto --max-num-seqs 16 --port 8000 --host 0.0.0.0 --enable-auto-tool-choice --tool-call-parser hermes
启动 openclaw
openclaw gateway --port 18789

夜雨聆风