Data Island Core

A versioned lakehouse engine with immutable storage, SQL analytics, enterprise security, and open table format interoperability. Your data, your infrastructure, your rules.

4 API Surfaces
5 Storage Backends
3 SQL Engines
16 Quality Checks
7 yr Audit Retention
OSS Source Available
Data Island Core — command center with SQL Editor, Tables, Ingestion and Data Quality
The Core command center: SQL, ingestion, quality, and governance in one place.

Core Capabilities

Nine pillars that make Data Island Core the foundation for modern data infrastructure.

Versioned Storage

Every write creates an immutable, append-only snapshot. Point-in-time queries, automatic deduplication, and soft-delete tombstones. No data is ever silently overwritten.

Learn more

Multi-Cloud Storage

Run on AWS S3, Azure Blob Storage, Google Cloud Storage, MinIO, or local disk. Switch backends with a single configuration change. Zero migration overhead.

Learn more

SQL Analytics

Standard SQL with automatic engine selection. DuckDB Lite for sub-second queries, DuckDB Pro with result caching, and Spark SQL via Thrift for distributed workloads.

Learn more

Security & Audit

Five-tier RBAC with row and column-level filtering. Tamper-evident SHA-256 hash chain audit logging with 21 structured fields. 7-year retention for DORA and SOC 2.

Learn more

OData for BI Tools

Dedicated OData 4.0 server for Power BI, Excel, and Tableau. Connect with a URL and bearer token — no drivers, no plugins, no ETL pipelines.

Learn more

MCP for AI Assistants

24 tools for Claude Desktop, Cursor, and any MCP-compatible client. Schema discovery, SQL execution, column profiling, and context persistence through natural conversation.

Learn more

Data Quality

16 built-in checks across type validation, completeness, and distribution analysis. 5 anomaly detectors, scheduled execution, and quality scores per table.

Learn more

Data Sharing

Cross-organization shares with zero-copy reads. Column-level and row-level filters, read-only access, bulk refresh, and on-demand materialization.

Learn more

Table Mirroring

Automatic export to Delta Lake, Apache Iceberg, and raw Parquet after every write. Integrate seamlessly with Databricks, Snowflake, and the Spark ecosystem.

Learn more

Architecture at a Glance

Four cooperating microservices, a Redis-backed catalog, and pluggable storage backends. Built to scale from a single laptop to a multi-region production cluster.

API Server

FastAPI core on port 8051. Owns SQL execution, RBAC, ingestion, sharing, mirroring, audit — exposed via REST + Python SDK.

Web UI

Jinja2 + HTMX admin and analyst interface on port 8050. SQL editor, table browser, quality dashboard, ingestion wizard, user management.

OData Server

Dedicated OData 4.0 service on port 8052 for Power BI, Excel, Tableau. No drivers, no plugins — just URL + bearer token.

MCP Server

Model Context Protocol on port 8099. 24 tools for Claude Desktop, Cursor, and any MCP-compatible AI assistant.

Storage layer is pluggable: local disk · S3 · Azure Blob · GCS · MinIO. Redis catalog tracks every snapshot. SQL routes to DuckDB Lite (sub-second), DuckDB Pro (cached), or Spark (distributed) automatically.

Security & Compliance

Five-tier RBAC, row- and column-level filters, tamper-evident audit, and 7-year retention. Built for DORA, SOC 2, and GDPR — not bolted on.

Five-Tier RBAC

viewer · analyst · editor · admin · superadmin. Per-table grants with row-level filter expressions and column masking. RBAC flows from Gatekeeper, not stored locally.

Tamper-Evident Audit

SHA-256 hash-chained audit log with 21 structured fields per event. Any tampering breaks the chain and is detected at verify-time.

Compliance-Ready

DORA, SOC 2, and GDPR controls baked in: encrypted at rest, 7-year retention, signed audit, data residency by storage backend.

Why Data Island Core?

See how Core compares to the alternatives across the dimensions that matter.

Capability Data Island Core Snowflake Databricks Raw Parquet
Source Available Partial
Versioned Storage Time Travel (90 days) Delta Lake
Multi-Cloud Vendor-managed Vendor-managed Manual
Self-Hosted Partial
SQL Engine DuckDB + Spark Proprietary Spark / Photon BYO Engine
RBAC Row + Column + Table
Audit Logging SHA-256 Hash Chain
OData for BI Tools
MCP / AI Integration Partial
Zero Vendor Lock-in

Start Building with Core

Explore every feature in detail or see which plan fits your needs.

Ready to Own Your Data?

Deploy Core on your infrastructure in minutes. Source-available, versioned, secure.