Architecture

DataNature spans the complete data journey as one governed flow through the open lakehouse. Every stage is governed end-to-end by DataNature Atlas — catalogued, secured, and lineage-tracked.

• Ingest — Connect, Stream, and DocuContext land data in the Bronze layer.
• Store & Engineer — the open lakehouse (Iceberg · medallion · vector store) and the dbt/Spark workbench refine data to Gold.
• Decide — DataNature AI and AI360 produce models, segments, and decisions.
• Consume & Activate — Voir agents, BI dashboards, and third-party apps put data to work.

Component Map

Component	Role
Connect / Stream / DocuContext	Batch, real-time, and document/audio ingestion into Bronze
Lakehouse	Open Iceberg storage, medallion layers, vector store, MPP OLAP + Trino federation
Data Engineering	dbt + Spark transformation workbench with Voir copilot
Lakebase (optional)	Managed Postgres OLTP serving layer for data & AI apps
AI / AI360	ML foundation; decision intelligence, Customer 360, semantic layer, reverse ETL, gateway
Voir	Multi-agent platform — Chat, Twins, Skills, Connectors, guardrails/HITL (web + mobile)
BI	Apache Superset dashboards + bring-your-own Power BI / Tableau
Atlas	Governance + semantic layer wrapping every component
Marketplace	Partner/ISV solutions run governed on-platform; become Voir agents
Platform / Lake / DevOps	Kubernetes runtime; optional managed object store; optional CI/CD & IaC

Table 2 · DataNature platform components and their roles.