Cost-Aware AI Gateway
kb:v1 • 0 docs redis + pg
Gateway Ready
Ask any question against your indexed PDFs and warmed FAQ cache. Every response updates the Live Execution Graph on the right.
Test Presets
Cost Saved vs Frontier
0.0%
$0.000000 saved
Cache Hit Rate
0.0%
0 requests
Avg Latency
0 ms
End-to-end gateway
Guardrails & Failovers
0 / 0
PII scrubs / 503 retries

Live Query Execution Graph

Preview Mode
HIT 0 LLM calls MISS EASY COMPLEX User Query POST /api/shop/chat 1. PII Ingress Guardrail Regex Scrub (<1ms) 2. Redis 8 Cache SHA-256 + Cosine <= 0.38 Redis Cache Hit Tier 1/2 • $0 LLM Spend 3. Redis 8 PDF RAG Top-4 Chunks + Citations 4. Query Router Heuristic + Centroid Gemini Flash Lite Fast Tier • $0.075/1M GPT-6.2-Sol Strong Tier • $2.00/1M 503 Auto-Failover + Write-Back 6. Async Postgres Telemetry BackgroundTasks (0ms block)
Click any node in the graph to inspect app/main.py

Send a message on the left to trace its exact live path, or click a node above to inspect its implementation.

Idle

Retrieved PDF Context

0 chunks
No RAG retrieval triggered yet.

PostgreSQL Request Logs

Cache Model Flags Latency Saved
Loading...