Home

VeriSimDB

Cross-system data consistency engine that catches drift before it causes damage.

Quick Start

See the main README for installation and usage.

cargo build --workspace                  # Build Rust core
cd elixir-orchestration && mix compile   # Build Elixir layer
cd elixir-orchestration && mix run ../demos/drift-detection/run_demo.exs  # Demo

Key Concepts

• Octad entity model -- Each entity exists simultaneously across 8 representations (Graph, Vector, Tensor, Semantic, Document, Temporal, Provenance, Spatial) with automatic drift detection and self-normalisation when any view diverges.
• Continuous drift detection -- When an entity's vector embedding diverges from its document content, or a provenance chain's hash integrity breaks, VeriSimDB identifies the drift type, scores severity, and triggers repair automatically -- before downstream failures occur.
• Dual-mode deployment -- Operates as a standalone database (Rust core with Oxigraph, HNSW, Tantivy, redb) or as a heterogeneous federation coordinator over existing databases (MongoDB, Redis, Neo4j, ClickHouse, SurrealDB, InfluxDB, MinIO).

Architecture

Rust workspace for performance-critical modality stores. Elixir/OTP orchestration layer for distributed coordination, GenServer-per-entity supervision, and drift monitoring. VQL (VeriSim Query Language) for native queries. Stapeln container ecosystem for verified deployment.

See docs/ for architecture details.

Related Projects

• PanLL -- Human-Things Interface with VeriSimDB dashboard panel
• Hypatia -- Neurosymbolic CI/CD (scan ingestion pipeline)
• Stapeln -- Visual container designer for verified deployment
• TypedQLiser -- Type-safe query layer
• proven -- Formal verification (certificate-based bridge)

Contributing

See CONTRIBUTING.md.

License

PMPL-1.0-or-later