Health Check GET /health — proxy liveness Comprehensive Health All services, DB, Redis List Models GET /v1/models — all providers Version Info Build, commit, feature flags

Basic Chat POST /v1/chat/completions (gpt-4o) Provider Prefix Routing anthropic/claude-sonnet-4-6 System Prompt System + user message Cache Hit Demo Same query → Phase 4 semantic hit

Claude Message POST /anthropic/v1/messages Claude Haiku Fastest Anthropic model

Cache Stats 4-phase hit rates + cost saved Skip Cache x-smartflow-cache: skip MCP Cache Stats Tool call result cache

Safe Query Policy passes — request proceeds PII Block SSN detected → 403 Forbidden Compliance Health ML engine status + model info

List Virtual Keys GET /api/enterprise/vkeys Create Virtual Key POST /api/enterprise/vkeys Check Budget GET /api/enterprise/vkeys/{id}/budget Budget Exceeded Virtual key → 429 over budget

🔧scenario mcp 🤖scenario a2a Cache Demo Miss → semantic hit (97x faster) Compliance Demo Safe → warn → block Virtual Key Lifecycle Create → use → budget → revoke Help / All Commands Full command reference