Getting Started

What is VectorScaleDB?

VectorScaleDB is a coupling intelligence engine built on a temporal-semantic database. The coupling matrix discovers cross-domain behavioral relationships that no other system can see — not by being told what to look for, but by learning what your data actually is.

Time-series data and vector embeddings share a unified proprietary index. This architecture enables queries impossible in any existing database: trajectory similarity search, behavioral compression, regime detection, cross-domain cascade prediction, and coupling traversal. Instead of stitching together a time-series database, a vector store, and a streaming engine, VectorScaleDB provides a single system with intelligence at the core.

Info

VectorScaleDB is designed for cross-domain behavioral intelligence, real-time entity tracking, anomaly detection, pattern matching, and coupling-driven prediction across a growing catalog of entity types spanning 20+ domains. Safety is enforced by mathematics, not policy.

Key Capabilities

Quick Start

1. Connect

The substrate is the VectorScaleDB network at https://api.vectorscaledb.com. Authenticate with an API key (or sign in with OAuth) — create one on the Authentication page, then pass it to the client. Running an air-gapped, standalone, or co-located node or cluster? Point the client at it directly instead — e.g. http://localhost:5434 for a VSDB or VSOS node on this device or local network.

Python 
from vectorscaledb import Client

client = Client("https://api.vectorscaledb.com", api_key="your-api-key")
curl 
# Verify connectivity
curl https://api.vectorscaledb.com/health

2. Ingest Data

Python 
client.ingest(
    entity_id="sensor-42",
    vector=[0.1, 0.8, 0.3, 0.9],
    entity_type="LINK_METRIC"
)
curl 
curl -X POST https://api.vectorscaledb.com/v1/ingest \
  -H "Authorization: Bearer $VSDB_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "entity_id": "sensor-42",
    "vector": [0.1, 0.8, 0.3, 0.9],
    "entity_type": "LINK_METRIC"
  }'

3. Query

Python 
results = client.temporal_knn(
    query_vector=[0.1, 0.7, 0.4, 0.8],
    start_time="2025-01-01T00:00:00Z",
    end_time="2025-12-31T23:59:59Z",
    k=5
)
curl 
curl -X POST https://api.vectorscaledb.com/v1/query/temporal-knn \
  -H "Authorization: Bearer $VSDB_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query_vector": [0.1, 0.7, 0.4, 0.8],
    "start_time": "2025-01-01T00:00:00Z",
    "end_time": "2025-12-31T23:59:59Z",
    "k": 5
  }'
Tip

Temporal KNN combines vector similarity with time-range filtering in a single index traversal -- no post-filtering required. This is what makes it faster than chaining a vector search with a time-series query.

Supported Protocols

Industry Adapters

VectorScaleDB includes adapters that translate domain-specific formats into temporal-semantic vectors on ingest:

Category Formats
Autonomous Vehicles MCAP, LAS/PCD, SPZ, glTF
Financial Markets FIX 4.4/5.0, CSV tick data, ITCH/OUCH
Networking & CDN NetFlow/sFlow/IPFIX, SNMP, gNMI, Prometheus
Bio-Computing NWB, HDF5, NEV/NSx, SpikeInterface, MEA streams, FlyWire connectome, FlyGM GNN, Brian2/NEST spike trains, MuJoCo behavior
IoT & Industrial MQTT (JSON/CBOR), Modbus, OPC-UA, CoAP
Film & VFX BVH, C3D motion capture, camera tracking
Gaming & NPC AI Game engine state, faction dynamics, social bonds
Geospatial ADS-B, AIS, TLE, weather, earthquake, fire, transit
Social Dynamics Social simulation, survey data
Genomics & DNA VCF variants, gene expression (bulk/single-cell), DNA foundation model embeddings
BCI & Cognitive BCI EEG signal frames, cognitive state decodings
Info

Adapters handle format parsing, vector extraction, and entity type mapping automatically. You send raw domain data; VectorScaleDB indexes it as temporal-semantic vectors.

Next Steps