Multi-model architecture

CrateDB is a multi-model database, meaning it can handle multiple data types,  all within one engine. Store relational, JSON, time-series, text, vector, and geospatial data side by side, and query across them seamlessly with SQL.

• Relational data: Traditional structured tables with full SQL joins and constraints.
• JSON & nested data: Semi-structured documents with flexible, evolving fields.
• Time-series data: High-ingest, high-cardinality datasets for real-time monitoring.
• Text & full-text search: Built-in indexing and scoring for fast keyword or semantic search.
• Geospatial data: Coordinates, shapes, and distance queries for location-based analytics.
• Vector data: Embeddings for AI-driven semantic and similarity search.

CrateDB’s unified model eliminates the need for multiple databases, simplifying architecture, improving performance, and reducing maintenance costs.

 

Dynamic schema

Data evolves and your database should, too. CrateDB’s dynamic schema feature allows new fields to be added automatically as data is ingested, without manual DDL changes or downtime.

• Flexible ingestion: Insert JSON or semi-structured data freely; CrateDB adapts automatically.
• Schema evolution: Add or modify columns dynamically as new attributes appear.
• SQL visibility: Newly discovered fields are instantly available through SQL queries and the system catalog.
• Use-case fit: Perfect for fast-moving environments like IoT, logs, or analytics pipelines.

This flexibility ensures developers can focus on building, not maintaining, while operations teams retain control through standard schema management and validation settings.

 

Automatic indexing

CrateDB’s automatic indexing system eliminates the need for manual index design or tuning. Every field, whether structured or semi-structured, is automatically indexed upon ingestion, enabling instant search and aggregations across any dataset.

• No manual setup: Indexes are created automatically for all supported data types.
• Optimized for performance: Built on Lucene for fast lookups, filtering, and sorting.
• Real-time search: Queries return results in milliseconds, even on millions of records.
• Unified with analytics: Use the same index for both SQL queries and full-text search.

This self-optimizing approach keeps your data instantly queryable,  removing the trade-off between flexibility and speed.

 

Why it matters

• Flexibility: Handle any data format — structured, semi-structured, or unstructured — without ETL complexity.
• Speed: Automatic indexing ensures sub-second query performance across growing datasets.
• Simplicity: Store and query all your data with SQL — no need for specialized APIs or connectors.
• Adaptability: Evolve schemas and workloads instantly as your business or data changes.
• Unified architecture: Avoid data silos and manage everything in one powerful, distributed engine.