VectorOWL

VectorOWL is MBSE discipline that survives AI-assisted development: one governed substrate so engineers, CI, and MCP-connected tools reason over the same versioned structure—OWL-backed graphs, vector retrieval over engineering data, Model Context Protocol integration (vectorowl-mcp), anchors where soft inference cannot override review obligations, and hooks aligned with computational-model characterization (trust and lifecycle metadata for simulations and models). This repository is the reference Rust implementation: a kernel-like runtime for graph–vector–constraint models (typed entities, deterministic anchors, hybrid queries, requirements and traceability, import bridges for CSV / RDF / XMI, RLVR environments, and the Axum HTTP surface vectorowld).

What VectorOWL is not: a new modeling language—you express the system in OWL (ontology); VectorOWL is the runtime and integration framework around that formal graph.

Two “MCP” phrases: Model Context Protocol = how hosts load vectorowl-mcp and sync tool context. Model Characterization Pattern (INCOSE community sense) = how enterprises describe trust and lifecycle for computational models—VectorOWL supports protocol wiring for integration and ontology-side hooks for characterization-style records.

Public narrative & install: Vector Stream Systems · Framework overview · MBSE overview & vectorowl-mcp install · Technical paper

Standards alignment (engineering volumes): INCOSE SE Handbook v4 · OMG UML 2.5.1 · OMG SysML v1.6 · OMG MOF 2.5.1 (see architecture volumes).

---

Architecture documentation

Document	Contents
docs/white-paper.md	VectorOWL + MCP: A Neuro-Symbolic Architecture for AI-Native Systems Engineering — conceptual whitepaper (neuro-symbolic stack, anchors, MCP, implementation summary)
docs/architecture/VOL-ARCH-001-kernel-runtime-architecture.md	Kernel (vectorowl crate): packages, class diagrams, sequences, state machines, activities, persistence, requirements traceability, glossary
docs/architecture/VOL-ARCH-002-system-integration-architecture.md	System integration: vectorowld, CLI, UI, HTTP catalog, deployment UML
docs/architecture/README.md	Index of architecture volumes
docs-export/vectorowl-query-language/README.md	VectorOWL Query Language (DSL, IQR, intent-based SIMILAR TO, execution model)
vectorowl-mcp-skill/README.md	Cursor Agent Skill (VectorOWL + MCP neuro-symbolic MBSE): install via scripts/install-vectorowl-skill.sh

Other references: MBSE_LANGUAGE_EVOLUTION.md, RL.md, docs/self-hosted-production.md.

---

Developer quick start (local)

Prerequisites: Rust toolchain (cargo). If cargo is not on your PATH, use e.g. $HOME/.cargo/bin/cargo.

1) Build

cargo check --workspace

2) Start API server (vectorowld)

cargo run -p vectorowld

Listens on http://0.0.0.0:8080 by default (VECTOROWL_HOST, VECTOROWL_PORT). Notable routes include POST /entity/create, POST /api/query, POST /rl/reset, GET /metrics, GET /openapi.json. Full catalog: VOL-ARCH-002 §10.

3) CLI (second terminal)

cargo run -p vectorowl-cli -- entity create --type Wing --attr area=20 --attr force=5000
cargo run -p vectorowl-cli -- query --type Wing
cargo run -p vectorowl-cli -- anchor create --type stress --target Wing
cargo run -p vectorowl-cli -- rl step --action update_area=30
cargo run -p vectorowl-cli -- rl reset

The CLI is a thin HTTP client; the server holds authoritative state.

4) Examples

cargo run --example car

5) Web UI

cd ui
npm install
npm run dev

The UI targets the API (default http://localhost:8080). For a single local URL via nginx, see docs/local-reverse-proxy.md and deploy/nginx/vectorowl-local.conf.

Deploy (API + CORS)

export VECTOROWL_HOST=0.0.0.0
export VECTOROWL_PORT=8080
export VECTOROWL_CORS_ALLOW_ORIGIN=https://your-site.com,https://www.your-site.com
cargo run -p vectorowld

UI production build:

cd ui
VECTOROWL_URL=https://vectorowl.vectorstreamsystems.com
VITE_VECTOROWL_API_URL=https://api.your-site.com npm run build

Self-hosted rollout: docs/self-hosted-production.md, deploy/deploy-prod.sh, deploy/systemd/vectorowld-prod.service, deploy/nginx/vectorowl-public.conf.

---

Workspace layout

Path	Role
vectorowl (root crate)	Library: McpRuntime, domain modules, api
vectorowld	Axum server binary
vectorowl-cli	HTTP CLI
ui/	Vite + React client
examples/	Runnable scenarios
mcp/	MCP tooling (see mcp/README.md)

---

Dependency manifest (root crate)

Crate	Purpose
uuid, serde, serde_json	Identity and JSON codecs
tokio	Async runtime, RwLock, broadcast
axum, tower-http	HTTP server and CORS
tracing, tracing-subscriber	Structured logs
csv, rio_*, roxmltree	Import bridges

Full table with versions: VOL-ARCH-001 Appendix B.