Platform Active

Orchestrate AI Agents
with Constitutional Governance

A multi-agent platform with built-in cost controls, model routing, and governance rules. Deploy intelligent agents that collaborate within defined boundaries.

5
Specialist Agents
6
LLM Models
13
Built-in Skills
90%
Quality Gate

Built for Production AI

Everything you need to run AI agents reliably, safely, and within budget.

Multi-Model Routing

Dynamically route tasks to the right model. Claude Opus for complex reasoning, Haiku for quick checks, with automatic fallback chains across providers.

Constitutional Governance

13-article governance framework with dual-approval workflows, quality gates, and violation escalation. Agents operate within defined boundaries.

Cost Controls

Per-agent daily ceilings, budget-aware model downgrading at 80% threshold, and real-time cost tracking with AUD/USD conversion.

Hub-and-Spoke Topology

Central orchestrator coordinates specialist agents. Research, development, operations, and content agents work in parallel with single-depth coordination.

Encrypted Communication

Agents communicate through NaCl-encrypted envelopes with replay protection. Ed25519 authentication with strict host key verification.

13 Built-in Skills

From deployment automation and security audits to email campaigns and legal document generation. A complete SaaS operations toolkit.

Designed for Control

Every decision is governed, every cost is tracked, every action is auditable.

# constitution.yaml
governance:
  topology: hub-and-spoke
  quality_gate: 90
  approval: dual (agent + human)
  budget: $10 AUD/day
# model_router.py
routing:
  primary: Claude Sonnet 4.5
  complex: Claude Opus 4.6
  fast: Claude Haiku 4.5
  fallback: [GPT-4.1, Gemini 2.5]
# violations → escalation
violations:
  MINORlog + continue
  MAJORalert + review
  CRITICALhalt + escalate

Constitutional AI Operations

Every agent operates under a machine-readable constitution that defines roles, limits, and escalation paths. No agent can exceed its authority.

  • Task-based model selection with keyword detection
  • Automatic Haiku fallback at 80% budget threshold
  • 90% quality gate enforced by Opus reviewer
  • Cross-provider fallback: Claude → GPT → Gemini
  • Full audit trail with 90-day retention

Three Agents, One Mission

A coordinated agent civilization where each agent has a distinct role and personality.

Nemo

Architect

The system architect running locally. Designs infrastructure, writes code, manages deployments, and maintains security posture.

Milo

Guardian

The VPS guardian running on cloud infrastructure. Manages services, handles backups, and communicates with the outside world.

OM

Quality Assurance

The testing specialist in the cloud. Runs end-to-end tests, validates deployments, and maintains the civilization dashboard.

Ready to Orchestrate?

Open source and ready to deploy. Start building your AI agent workforce today.