Skip to main content

Getting Started with ScramDB

ScramDB is an adaptive HTAP SQL database that compiles queries to native machine code. It speaks the PostgreSQL wire protocol - connect with psql or any PostgreSQL driver.

Why ScramDB?

Most databases force you to choose: fast analytics or easy SQL. ScramDB gives you both.

Traditional OLAPScramDB
Query languageProprietary or limited SQLFull PostgreSQL-compatible SQL
Query executionInterpreted or vectorizedJIT-compiled to native machine code
ParallelismThread poolCore-pinned morsel-driven (linear scaling)
StorageRow-store or external filesPurpose-built columnar engine (Tundra)
Vector searchSeparate systemBuilt-in pgvector-compatible
GPU accelerationNot availableBuilt-in for NVIDIA, AMD, Apple Metal
Client compatibilityCustom drivers neededAny PostgreSQL driver works
TransactionsOften missingFull MVCC with serializable isolation

No new query language. No proprietary SDK. No vendor lock-in. Just connect with psql and run SQL.


What is ScramDB?

ScramDB is a high-performance analytical database written in Rust, designed for:

  • OLAP analytics - Sub-second complex queries over millions of rows
  • HTAP workloads - Mix transactional and analytical queries on the same data
  • Vector search - pgvector-compatible similarity search
  • Real-time dashboards - Streaming query results with backpressure

Key Capabilities

FeatureDetails
Query executionJIT compilation, morsel-driven parallelism
SQL support72+ functions, window functions, CTEs, subqueries
StorageTundra columnar engine with zone maps
TransactionsMVCC (Read Committed → Serializable)
ProtocolPostgreSQL wire protocol
IndexingART, B+Tree, Hash
Vector searchpgvector-compatible VECTOR(n) type

🚀 Quick Start

Get running with Docker in 2 minutes

📚 SQL Reference

Complete SQL compatibility reference

Next Steps

  1. Quick Start - Docker setup and first connection
  2. First Queries - Create tables and run SQL
  3. SQL Reference - Full SQL compatibility
  4. Benchmarks - TPC-H performance results