Skip to content
Vector Stream Systems logo Vector Stream Systems

What you'll see

A full VectorOWL instance, pre-seeded and ready

Entity graph

Browse a pre-loaded systems graph — nodes, edges, and OWL-classified entities. Inspect attributes, follow links, and visualize topology in 3D.

OWL reasoning

Run SPARQL-like queries over the ontology. Trace subclass relationships, check satisfiability, and see how formal axioms constrain the model.

Vector search

Search by meaning, not keyword. Vector embeddings rank similar designs, requirements, and telemetry by semantic proximity.

Anchor constraints

Evaluate hard constraints that override probabilistic suggestions. See pass/fail results with severity levels and audit trails.

Live demo

Hosted instance — early access

A hosted VectorOWL demo is being prepared. Request early access and you'll receive a login link when your slot is ready.

GitHub Sponsors get priority access — sponsors are added to the demo within 24 hours.

Hybrid reasoning blend (illustrative)

VectorOWL mixes symbolic OWL constraints with embedding similarity. This slider does not call a server — it only illustrates how a tunable weight α biases the response toward symbolic versus vector pathways.

Run locally

Prefer to run VectorOWL on your own machine? Build the Rust backend and the React UI from source.

Show terminal instructions
bash
# Terminal 1 — start the backend
cd VectorOWL
export LD_LIBRARY_PATH="target/release/build/torch-sys-*/out/libtorch/libtorch/lib:$LD_LIBRARY_PATH"
VECTOROWL_REQUIRE_TORCH_GPU=false ./target/release/vectorowld

# Terminal 2 — start the UI
cd VectorOWL/ui
npm run preview -- --host

# Open http://localhost:5173

Requires Rust, Node 22, and a LibTorch setup. Or use Docker: docker-compose up --build in the VectorOWL repo.

API status
Click to check API reachability.