Documentation

Stellaris AOS Documentation

Architecture, API reference, deployment models, and governance frameworks for enterprise agentic orchestration.

Core Architecture Concepts

Stellaris AOS is designed as a secure, deterministic routing layer between foundation models (LLMs) and enterprise infrastructure. It acts as the Agentic Operating System that governs how, when, and why an AI agent can execute an action.

Deterministic Routing

Ensures that agent intents are mapped to strictly typed API endpoints with predefined schemas, eliminating hallucinated function calls.

Governance Layer

Every mutation request passes through the HITL (Human-In-The-Loop) validation gateway before execution.

Human-In-The-Loop (HITL)

For high-risk operations, Stellaris AOS pauses execution and queues a validation request for a human operator.

[Agent: Procurement] -> Intent: Purchase Order [Stellaris AOS] -> Intercepted: High Risk (Budget > $0) [Stellaris AOS] -> State: PAUSED_WAITING_APPROVAL [Human Operator] -> Validates via Dashboard -> State: APPROVED [Stellaris AOS] -> Executes API Call -> SAP / ERP

Competitor Benchmarking

CapabilityStellaris AOSStandard iPaaS (e.g., Zapier)Custom Python Scripts
Multi-Agent State SyncNative (Graph-based)NoManual Dev Required
HITL ConstraintsEnterprise GradeBasic (Email approvals)Manual Dev Required
ISO 42001 AuditingBuilt-in (Immutable Logs)NoneNone