Skip to main content

Welcome to RavenmaskOS

RavenmaskOS is an AI agent platform that provides enterprise-grade infrastructure for building, deploying, and managing autonomous AI agents at scale.

What is RavenmaskOS?​

RavenmaskOS is built on a cognitive architecture inspired by Norse mythology, designed to give AI agents memory, reasoning, and the ability to learn from experience. The platform handles the complex infrastructure so you can focus on building intelligent automation.

Key Capabilities​

  • Cognitive Architecture: The Raven Model (Huginn for reasoning, Muninn for memory) provides a structured approach to agent intelligence
  • Graph Memory: Neo4j-powered relationship memory that helps agents understand context and connections
  • 400+ Integrations: Deep n8n integration brings enterprise SaaS tools as first-class platform components
  • Human-in-the-Loop: Built-in approval workflows, case management, and audit trails for production safety
  • Multi-tenant: Enterprise-ready with user management, RBAC, and workspace isolation

Architecture Overview​

┌─────────────────────────────────────────────────────────────────┐
│                       RavenmaskOS Platform                       │
├─────────────────────────────────────────────────────────────────┤
│                                                                   │
│  ┌──────────────┐      ┌──────────────┐      ┌──────────────┐  │
│  │   Norns      │      │   Bifrost    │      │     HITL     │  │
│  │   Studio     │─────▶│     API      │◀─────│   System     │  │
│  │  (Builder)   │      │   Gateway    │      │  (Approval)  │  │
│  └──────────────┘      └──────┬───────┘      └──────────────┘  │
│                               │                                  │
│  ┌──────────────┐      ┌──────┴───────┐      ┌──────────────┐  │
│  │   Huginn     │      │    Domain    │      │    n8n       │  │
│  │  (Thought)   │◀─────│ Intelligence │─────▶│ Integration  │  │
│  │   LangGraph  │      │   Schema     │      │   Engine     │  │
│  └──────┬───────┘      └──────────────┘      └──────────────┘  │
│         │                                                        │
│  ┌──────┴───────┐      ┌──────────────┐      ┌──────────────┐  │
│  │   Muninn     │      │  PostgreSQL  │      │     NATS     │  │
│  │  (Memory)    │◀─────│   (Source    │─────▶│  JetStream   │  │
│  │   Neo4j      │      │  of Truth)   │      │   (Events)   │  │
│  └──────────────┘      └──────────────┘      └──────────────┘  │
│                                                                   │
│  ┌─────────────────────────────────────────────────────────┐   │
│  │              Langfuse (Observability)                    │   │
│  └─────────────────────────────────────────────────────────┘   │
│                                                                   │
└─────────────────────────────────────────────────────────────────┘

Core Components​

The Raven Model​

Our cognitive architecture is inspired by Odin's ravens from Norse mythology:

  • Huginn (Thought): Strategic reasoning, planning, and decision-making powered by LangGraph
  • Muninn (Memory): Neo4j graph database for relationship memory plus PostgreSQL for structured data
  • Context Window: Working memory for active agent sessions
  • Domain Layer: Pluggable schemas for different business domains (finance, health, home automation, etc.)

Bifrost API Gateway​

All agent tools and integrations flow through Bifrost, providing:

  • Unified API interface for all external services
  • Rate limiting and quota management
  • Authentication and authorization
  • Request/response logging and audit trails

HITL (Human-in-the-Loop)​

Production-ready governance for AI agents:

  • Approval workflows for sensitive operations
  • Case management for agent escalations
  • Audit trails for compliance
  • User feedback loops for continuous improvement

Domain Intelligence Schema (DIS)​

Formal schema system for modeling business domains:

  • Type-safe domain definitions
  • Relationship mapping between entities
  • Business rule validation
  • Cross-domain data sharing

Technology Stack​

  • Agent Orchestration: LangGraph
  • Databases: PostgreSQL (source of truth) + Neo4j (graph memory)
  • Messaging: NATS JetStream for real-time events
  • Integrations: n8n (400+ pre-built connectors)
  • Observability: Langfuse for LLM traces and metrics
  • Reverse Proxy: Traefik with automatic SSL

Next Steps​

Ready to build your first AI agent? Start with our Quickstart Guide to create and deploy an agent in minutes.