DeltaForge is a SQL engine for Delta Lake and Apache Iceberg. It installs on your own infrastructure: cloud VMs, on-premises servers, or air-gapped environments. No JVM, no external runtime dependencies. Every feature below ships on every plan.
The foundational engines and services that power everything
Extended SQL engine. Familiar PostgreSQL syntax plus PIPELINE, VACUUM, OPTIMIZE, time travel, and Delta Lake & Iceberg commands built in.
Learn more →Power BI, Tableau, Excel, Python, .NET, and R connect to Delta tables directly via a native ODBC driver. No secondary copy required.
Learn more →ADBC driver for Delta Lake with Power BI Desktop integration. Apache Arrow record batches handed to the BI tool by reference. 13.7x faster than .NET ODBC on a 1M-row scan.
Learn more →Vectorized, SIMD-accelerated query execution. Parallel task scheduling, memory management, streaming results. Transparent DFCU metering: pay for query execution only, no cluster idle costs.
Learn more →Unified catalog, RBAC, credential management, audit logging, workspace isolation.
Learn more →Capacity follows demand. Add workers when queries arrive, scale to zero when idle, keep a warm baseline for instant response. Runs in your own cloud.
Learn more →Common file formats, relational databases (PostgreSQL, MySQL, SQL Server, Oracle), cloud storage. Visual flattener turns complex nested formats (JSON, XML, EDI, HL7, FHIR, Protobuf) into flat SQL tables. Schema inference, predicate pushdown, parallel ingestion.
Learn more →First-class support for the two leading open table formats
Native Delta Lake engine. ACID transactions, time travel, schema evolution, deletion vectors, CDC, predictive optimization. No managed runtime required.
Learn more →V1/V2/V3 feature coverage. Native reads, UniForm interoperability, schema evolution, deletion vectors, and open-format workflows.
Learn more →Optional PGM, B+ tree, and bloom filter indexes for Delta tables. Used by the DeltaForge planner to skip files and locate rows. Full CREATE INDEX DDL.
Advanced analytical capabilities built directly into the engine
Zero-ETL graph projections over the Delta tables you already run. Cypher plus 32 algorithms including PageRank, Leiden, Bellman-Ford, FastRP embeddings, K-core, and node similarity. Eighteen of them run on GPU via the ON GPU hint.
Uber H3 hexagonal indexing. Spatial aggregation, distance calculations, hierarchical resolution.
Learn more →200+ built-in functions including H3 geospatial. Scalar, aggregate, and table-valued. Custom UDFs in SQL.
Learn more →CREATE CHART wraps a query and returns SVG. Ten chart types: BAR, HBAR, LINE, AREA, SCATTER, PIE, HISTOGRAM, HEATMAP, RADAR, CANDLESTICK.
Native parsing for healthcare, supply chain, and data protection standards
Native FHIR R4/R5 parsing. JSON, XML, NDJSON formats. Query clinical resources with SQL.
Learn more →Zero-copy HL7v2 message parsing. SIMD-accelerated segment extraction.
Learn more →X12, EDIFACT, TRADACOMS parsing. Transaction sets queryable as SQL tables.
Learn more →Pseudonymisation controls for GDPR-aligned workflows. Hash, encrypt, mask, and redact selected fields. Reversible options are available where configured.
Learn more →Tools and protocols that integrate with your existing workflow
Desktop GUI (Tauri), VS Code extension, CLI. Query editor with IntelliSense, visual explain plans, result streaming.
Learn more →Pipeline code lives in Git. Branch, develop, commit, auto-deploy. DeltaForge scans your repository, extracts lineage, computes execution order, and schedules runs. No DAGs, no YAML. Just SQL.
Learn more →Built-in MCP server (the open Model Context Protocol used by Claude, Cursor, and Copilot Chat). Typed actions for catalog, lineage, SQL, and pipelines. Scoped tokens, RBAC enforcement, and tool-call audit logging. Real parser feedback helps improve AI-generated SQL.
Learn more →Security controls and deployment flexibility for production environments
Common questions about running SQL on Delta Lake and Iceberg with DeltaForge
Yes. DeltaForge executes MERGE, UPDATE, and DELETE on Delta Lake tables natively, writing the Delta transaction log directly on your storage. There is no Spark cluster, no JVM, and no managed runtime involved. See the Delta Lake tables page for the full write surface.
Yes. The same engine covers Apache Iceberg V1, V2, and V3 features, including native reads, UniForm interoperability, schema evolution, and deletion vectors, so Delta and Iceberg tables sit side by side in one catalog.
Yes. The bundled ODBC driver connects Power BI, Tableau, Excel, .NET, Python, and R to Delta tables on your storage. The bundled ADBC driver plus Power Query connector hands Arrow record batches to Power BI Desktop by reference.
Every plan ships the same engine, the same drivers, the same Cypher graph layer, the same MCP server, and the same control plane. Plans differ on scale, included compute units, and support; they do not gate features. See pricing for details.
Hands-on guides for the features above, written against a running DeltaForge install
There are no add-ons and no paywalled features. Plans differ only by capacity and included DFCU.