Autonomous · AI-native · No build required

Autonomous Systems Engineering Platform

Four interlocking layers that draft requirements, synthesize complex safety assessments, reason about architecture, propagate change impact, manage verification evidence, and synchronize your engineering toolchain. Engineers review and approve. One runtime. Governed by default.

View on GitHub Read the White Paper

System Architecture

┌─────────────────────────────────────────────────────────────┐
│         CAD / CAE / PLM / MBSE & Req. Mgmt Tools            │
│   (CATIA, Ansys, MATLAB, Cameo MBSE, DOORS, JAMA, custom)   │
└──────────────────────┬──────────────────────────────────────┘
                       │ Model Context Protocol
                       ▼
┌─────────────────────────────────────────────────────────────┐
│  ┌─────────────┐  ┌─────────────┐  ┌─────────────┐          │
│  │   Ontology  │  │   Vector    │  │     MCP     │          │
│  │    Layer    │  │    Layer    │  │Coordination │          │
│  │  OWL / RDF  │  │ HNSW / Faiss│  │IdentityReg  │          │
│  │  SPARQL     │  │  Embeddings │  │ContextUpdate│          │
│  └─────────────┘  └─────────────┘  └─────────────┘          │
│         \                │                /                 │
│          \               │               /                  │
│           ▼              ▼              ▼                   │
│               ┌─────────────────────────────┐               │
│               │      Hybrid Reasoning       │               │
│               │   α · Symbolic + (1−α) ·    │               │
│               │      Vector Similarity      │               │
│               └─────────────────────────────┘               │
│                          │                                  │
│                          ▼                                  │
│               ┌─────────────────────────────┐               │
│               │   IdentityRegistry + DAG    │               │
│               │   ContextUpdate Events      │               │
│               └─────────────────────────────┘               │
└─────────────────────────┬───────────────────────────────────┘
                          │ Model Context Protocol
                          ▼
┌─────────────────────────────────────────────────────────────┐
│         AI Tools  (Claude, Cursor, Copilot, agents)         │
└─────────────────────────────────────────────────────────────┘

System introspection

Live system reports at a glance

VectorMBE generates real-time reports on entity counts, graph topology, and requirement coverage with no manual querying required.

Features

Four layers, one substrate

Ontology Layer

OWL/RDF triple store with SPARQL access. Imports SysML v1/v2, UML 2.x, and Ecore/XMI natively. Blocks, requirements, state machines, allocations, and flows become graph nodes and edges automatically.

Vector Layer

HNSW approximate nearest-neighbor search over high-dimensional embeddings. Similarity queries over requirements, design entities, and simulation results.

MCP Coordination

Model Context Protocol servers at tool boundaries. ContextUpdate events propagate through a dependency DAG with an IdentityRegistry.

Natural Language Interface

Ask the graph in plain English, such as "Which functions satisfy requirement R-042?" and get traceable subgraphs back. LLM-powered GraphRAG grounds every answer in the ontology.

API Reference

REST API with OpenAPI documentation. Entity CRUD, vector queries, and graph traversal.

View API Docs

Self-Hosted

Run on your own infrastructure. Single binary. No external dependencies except a triple store and optional GPU for embeddings.

Install Guide

Cloud

Managed VectorMBE hosted by us. SSO, audit logging, and SLA-backed support for teams that need zero ops overhead.

View Pricing