Technical Architecture

• Terminology
• Threading model
• Listener
  • Downstream (Ingress)
  • Upstream (Egress)
  • Configure Listener
• Prompts
• Messages
• Prompt Guardrails
• Prompt Targets
  • Intent Detection and Prompt Matching:
  • Agentic Apps via Prompt Targets
• Prompting LLMs
  • Example: Using the Arch Python SDK
  • Example: Using OpenAI Client with Arch as an Egress Gateway
• LLM Provider
• Example: Using the Arch Python SDK
• Model Serving
  • Local Serving (CPU - Moderate)
  • Local Serving (GPU- Fast)
  • Cloud Serving (GPU - Blazing Fast)