Demo Applications: Four Industrial Systems. One real-time database

Real-time process monitoring for a large-scale brewery with approximately 500 employees across 238 sensors in 13 production zones, with AI anomaly detection and a live process digital twin.

This demo shows how CrateDB can serve as a unified database for storing, analysing, and visualising IoT telemetry data in real time instead of a common but complex multi-database architecture.

It highlights day-based automatic partitioning for fast queries and easy retention, flexible columns that let AI results and sensor metadata evolve without schema migrations, and in-database SQL analytics (window functions, aggregations, time filtering) fast enough to replace a separate OLAP system entirely.

🗎 Read the docs

This demo simulates a smart factory with CNC machines, industrial robots, and autonomous guided vehicles, collecting sensor data in real time and providing operators with a unified view of production status.

You can see live operations monitoring and an AGV floor map, interactive time-series analysis with window functions, AI-driven predictive maintenance via vector similarity, semantic search across maintenance docs (full-text + vector), and a 3D factory digital twin.

Explore how a single CrateDB database can replace an entire stack of systems. Instead of maintaining separate databases for time series, geospatial data, full-text search, and vector similarity, everything resides in one cluster with a single SQL interface.

🗎 Read the docs

This CrateDB demo shows an entire Network Operations Center for 15,700 households with ~31,000 meters in the city of Zagreb. Explore real-time telemetry ingestion at scale, anomaly and leak detection via AI, predictive maintenance scheduling, GIS-based network visualization, and operational dashboards.

It showcases CrateDB's ability to handle both high-throughput OLTP-style ingestion (2,500+ readings/sec) and OLAP-style aggregations across 2.5M+ rows in sub-second latency simultaneously. No ETL pipeline or separate data warehouse needed.

🗎 Read the docs

This demo models a small hydro plant with two Francis turbine units, simulating 83 sensor tags across electrical, mechanical, and PLC systems, and uses CrateDB as the central datastore behind a live SCADA digital twin.

Explore an animated process diagram with live gauges and trends, a multi-level alarm engine with acknowledgement workflows, AI-based anomaly detection that flags the most-deviating sensor, and predictive maintenance with remaining-useful-life estimates, plus a "what-if" mode to inject faults and watch the system react.

🗎 Read the docs

Millions of events per second flow in over MQTT and Socket.IO — queryable the instant they land.

Time-series, JSON, full-text, vector and geospatial data all answered in a single standard SQL query. 

IsolationForest anomaly detection and RUL prognostics run against data that just arrived — no batch job. 

3D digital twins and geospatial maps redraw continuously — driven straight off the query results. 

JOINs, aggregations and window functions over billions of rows — no pre-aggregation, no schema redesign.

Time-series, JSON, full-text, vector, geospatial and structured data — all in one engine.

Ingest millions of events per second from IoT sensors, Kafka or logs — immediately queryable.

Columnar storage and parallel execution return results in milliseconds on production workloads.

Add nodes to linearly increase throughput. Automatic sharding and replication, no manual work.

Apache 2.0 licensed. Run it on-prem, at the edge, in the cloud, or fully managed. No lock-in.