← Back to Services
🐉

Taniwha Gateway

Constitutional AI Orchestration Layer

Taniwha v4.0 (Phase 3) • Live health + query + metrics

Production

The Infrastructure Behind All Services

Taniwha Gateway is the unified orchestration layer powering all Axiom Intelligence services. Every request flows through constitutional filtering, Scrooge cost optimization, and provider speed routing before reaching an LLM.

In our demo, a 7-year-old requests a space story through House of Adventure. The Gateway validates constitutional safety (Brahmacharya 0.97 for child-appropriate content), routes to the fastest Layer 1 provider (Claude Sonnet, 780ms), and delivers the story for $0.002 instead of $0.35 direct cost—a 175× cost reduction.

🎯 6-Layer IAMP Architecture

L0 • Deterministic

$0.00

Pattern matching, regex • Zero cost

L1 • Cache

$0.00

Semantic similarity • Zero cost

L2 • Workers AI

$0.033

Llama 3.2 3B • Simple queries

L3 • Scout

~$0.002

Claude Haiku • Fast responses

L4 • Warrior

~$0.01

Claude Sonnet • Complex tasks

L5 • Elder

~$0.018

Claude Opus 4 • Deep reasoning

Volume-based routing: Query complexity automatically selects the optimal layer based on Five Yamas scores and log depth.

Key Features

  • ✓ Constitutional filtering (Five Yamas)
  • ✓ Provider speed routing (sub-10ms)
  • ✓ Federation token security
  • ✓ Rate limiting (50 req/10s burst)
  • ✓ WebSocket admin interface
Live Status
Checking Taniwha…

Metrics

Requests (24h)
~150
Average Latency
2.5s
Cost Reduction
96.5%
Constitutional Safety
100% ✅
Watch Interactive Demo → Admin Live Console (Dr Sage) → Request API Access