Skip to main content

​
Deploy icon Deploy VPS CPU Hybrid

Launch profile:
  • 8 vCPU
  • 24 GB RAM
  • NVMe storage
  • CPU-only local inference + remote fallback

​
Architecture

  1. AiCord bot calls api.aicord.cloud
  2. aicord/fast routes to local CPU model
  3. Heavy or overflow requests route to remote upstream
REQUEST_TIMEOUT_MS=12000
RATE_LIMIT_PER_MINUTE=120
ALLOW_DIRECT_MODEL=false
UPSTREAM_OVERRIDE_MODEL=qwen2.5-coder:7b

​
Deploy checklist

  • Provision VPS and update packages
  • Install Python 3.11+ and create virtual environment
  • Configure AICORDCLOUD/.env
  • Start gateway (python run.py or process manager)
  • Configure reverse proxy with TLS
  • Point api.aicord.cloud DNS to proxy