AI Model Gateway

Acknowledgement: Shared by Peter Bamuhigire, techguypeter.com, +256 784 464178.

Use When

• Designing or implementing the LLM gateway as a control-plane service for a multi-tenant SaaS.
• Replacing direct OpenAI / Anthropic / Bedrock SDK calls in feature code with a routed, audited, cost-attributed path.
• Adding multi-provider fallback to an existing AI feature (Anthropic → Bedrock → OpenAI).
• Enforcing per-tenant token caps, regional routing, and audit at one chokepoint instead of scattered library calls.

Do Not Use When

• The task is the wider AI architecture — start with ai-on-saas-architecture.
• The task is direct provider API exploration / spike — ai-llm-integration is the bare-metal SDK skill.
• The task is the prompt design — use ai-prompt-engineering.

Required Inputs