Summary Saiku Analytics offers an open-source OLAP browser and data querying solution for engineers and power users. It provides drag-and-drop cube analysis, SQL access via Mondrian/Calcite, and a REST API for AI agents to query data without MDX knowledge.

How It Works It uses Mondrian 4.8.1.x with Calcite-based SQL planner (default). Data results use Apache Arrow for zero-copy transfer. Backend: Jetty 12, Spring 6, Jersey 3.1; Frontend: SvelteKit 5 SPA. Its typed REST API for LLM agents translates natural language to MDX with structured, self-correcting error feedback.

Quick Start & Requirements Docker quick start: docker run -d -p 8080:8080 --name saiku ghcr.io/spiculedata/saiku, includes H2 demo cube. Source build requires JDK 21, Maven 3.9+. Fat JAR: mvn -pl saiku-launcher -am -Dmaven.test.skip=true package. Live demo: https://demo.saiku.bi.

Highlighted Details

• AI Query API: Typed REST surface (/rest/saiku/api/ai/*) enables AI agents to discover schemas and query data, translating natural language to MDX with structured, self-correcting error feedback.
• Apache Arrow: Integrates Arrow wire format for cellsets, enabling zero-copy data transfer