Getting Started

VioGraph turns your data warehouse into a graph analytics layer — query Snowflake, BigQuery, Databricks, or Starburst as a graph, 120,000× faster than SQL joins, with zero ETL and zero idle cost.

What’s in this section

• Quickstart — connect a data source, map your tables, launch a graph, and run your first Cypher query in under 10 minutes.
• Architecture — how the control plane, wrappers, and export worker fit together.
• Deployment — install on Kubernetes (GKE, EKS, AKS, or on-prem).

Just exploring?

The public demo at ai-graph-olap.com/dashboard is pre-loaded with a fraud-detection dataset you can query immediately. No signup required.

Two-minute overview

1. Connect — point VioGraph at your warehouse using existing credentials.
2. Map — decide which tables become nodes (e.g. Account, Customer) and which become relationships (e.g. TRANSFERRED_TO).
3. Launch — the control plane spins up an ephemeral graph instance, loads your data, and is ready to query.
4. Query — use Cypher from the dashboard, the Python SDK, or the AI assistant. Results in milliseconds.

When you stop querying, the graph tears itself down. You only pay for compute while actively using it.