Architecture

The coupling matrix is the product. The database is the substrate. The network is the business model. A layered Rust engine with coupling intelligence at the core, safety enforced by mathematics, and trust-verified federation at the edge.

System Architecture

Eleven architectural layers from API surface to trust-verified federation, with the coupling matrix as the central intelligence. hundreds of crates, universal binary with hardware acceleration, adaptive deployment from microcontroller-class to datacenter, zero warnings.

API Layer — 450+ Endpoints
REST gRPC GraphQL SSE Streaming WebSocket Subscriptions Webhooks MCP (131 AI Agent Tools)
Middleware & Security
Auth / RBAC Rate Limiting Audit Chain Multi-Tenant Cloud Platform Envelope Encryption Differential Privacy WASM Reducers Bilateral Consent
Coupling Matrix — The Central Intelligence
7 Coupling Subsystems Drift-Correlation Discovery Cross-Entity Prediction Coupling-Enhanced Compression Behavioral Regime Tracking Cross-Domain Cascade Detection Data Discovery
Trust Network — Safety as Physics
3-of-5 Threshold Signing Behavioral Trust Scoring Anti-Impersonation No-Single-Entity Control Irrevocable Consent Gate Safety Bitmap Attestation Revocation Propagation
Core Engine — 200+ Entity Types, 20+ Domains
Ingestion
One convergence pipeline. Any of 36+ formats — robotics logs, financial feeds, genomic variants, 3D scenes, neural recordings — is detected, normalized to a unified representation, and handed to the same compression and index engine. Add a format without touching the storage or query path.
Compression
200–16,500x on vectors via multi-stage behavioral compression. 30–500x on non-vector data. Cross-entity correlation for additional 2–5x on correlated clusters
Temporal-Semantic Index
Proprietary temporal-semantic index, sub-ms KNN, trajectory similarity, multi-resolution hierarchy, snapshot/restore
Query Engine
100+ operations, anomaly scoring, pattern matching, forecasting, density-based behavioral clustering, cascade detection, genomics analysis, neuroscience-inspired pattern separation
Unified Storage
Vectors + Documents + Graphs + Blobs + Full-Text Tiered Storage (Hot / Warm / Cold) Four-Frequency Data Lifecycle Semantic Chunk Deduplication Snapshot / Restore
Distribution — Adaptive Deployment (Microcontroller to Datacenter) · Fully-Federated / Lightly-Coupled / Air-Gapped Modes
Consistent Hashing Membership Gossip Conflict-Free Replication Content-Addressed Archival P2P Transport Federated Templates Dynamic Hierarchy Overlay Hub Topology Island Mode

Quick Start

Ingest temporal-semantic data in three lines. SDKs for Python, Rust, and raw HTTP.

# pip install vectorscaledb  (SDK available at launch)
from vectorscaledb import Client

client = Client("https://api.vectorscaledb.com")

# Ingest a temporal vector
client.ingest({
    "entity_id": "drone-7",
    "timestamp": "2026-03-08T12:00:00Z",
    "vector": [0.1, 0.9, 0.3, 0.7],
    "metadata": {"altitude": 120.5}
})

# Find similar trajectories
results = client.query.trajectory_similarity(
    entity_id="drone-7",
    time_range=("2026-03-07", "2026-03-08"),
    top_k=5
)
use vectorscaledb::Client;

let client = Client::new("https://api.vectorscaledb.com");

// Ingest a temporal vector
client.ingest(IngestRequest {
    entity_id: "drone-7".into(),
    timestamp: Utc::now(),
    vector: vec![0.1, 0.9, 0.3, 0.7],
    metadata: json!({ "altitude": 120.5 }),
}).await?;

// Find similar trajectories
let results = client.trajectory_similarity(
    "drone-7",
    time_range,
    5,
).await?;
# Ingest a temporal vector
curl -X POST https://api.vectorscaledb.com/api/v1/ingest \
  -H "Content-Type: application/json" \
  -d '{
    "entity_id": "drone-7",
    "timestamp": "2026-03-08T12:00:00Z",
    "vector": [0.1, 0.9, 0.3, 0.7],
    "metadata": {"altitude": 120.5}
  }'

# Find similar trajectories
curl -X POST https://api.vectorscaledb.com/api/v1/query/trajectory-similarity \
  -H "Content-Type: application/json" \
  -d '{
    "entity_id": "drone-7",
    "start": "2026-03-07T00:00:00Z",
    "end": "2026-03-08T00:00:00Z",
    "top_k": 5
  }'

API Preview

Key endpoints across 450+ total. Ingestion, query, coupling, and management surfaces.

Method Endpoint Description
POST /api/v1/ingest Ingest a temporal vector with metadata
POST /api/v1/ingest/batch Batch ingest up to 10,000 vectors
POST /api/v1/query/knn K-nearest neighbor search with time filter
POST /api/v1/query/trajectory-similarity Find entities with similar behavioral trajectories
POST /api/v1/query/anomaly-score Score entity behavior against baseline
POST /api/v1/query/pattern-match Match behavioral patterns across entities
GET /api/v1/entity/{id}/history Full temporal history for an entity
GET /api/v1/compression/stats Compression ratios and regime counts
GET /api/v1/health Health check with storage connectivity
Full API Reference →

Ready to see coupling intelligence in action?

The coupling matrix discovers what your data knows. See it work with your domains.

Request Demo