DeltaForge

DeltaForge is a native lakehouse engine for Delta Lake and Apache Iceberg. It runs both formats in a single process on local hardware and eliminates the need for Spark clusters, data duplication, or a second warehouse.

The engine supports reading and writing with first-class grammar for MERGE, time travel, deletion vectors, UniForm interop, and change data feed. Correctness is established through two independent verification layers that derive every expected value outside the engine under test. Downloadable use cases execute on the installed engine, with every failure self-diagnosing. It maintains bi-directional conformance against Apache Spark, publishing schemas, sample rows, and verification arithmetic for open inspection. Example checks include uniform_v3_roundtrip returning identical numeric results and merge_dv_concurrent producing matching row counts.

BI tools connect directly via ADBC and ODBC. Graph workloads run with Cypher on the same data. An MCP server supplies typed catalog access to AI agents. Delivery occurs through native binaries available via winget, Chocolatey, and the Azure Marketplace. The project maintains a GitHub repository.

A Community License is available at no cost with no credit card required. Paid options appear under a separate pricing page.