BigHugger
GH Repository · spiculedata

saiku

Open-source semantic layer: one cube for Excel (MDX/XMLA), dashboards, and AI agents (MCP). Mondrian + Apache Calcite.

stars
1,324
30-day movement
+10/day
Related entries
60
Connections
2
mcp-serverdockermdxJavamcpjava/mavenbusiness-intelligencesemantic-layerapache-calcitexmlaolapmondriananalytics

Saiku is an open-source semantic layer built on Mondrian and Apache Calcite that exposes a single OLAP cube via MDX/XMLA. It serves traditional BI clients like Excel and dashboards as well as AI agents through an MCP server interface.

It lets you define one semantic model once and query it from Excel, dashboards, and MCP-capable AI agents.

Use it to

  • Query the semantic layer from Excel via MDX/XMLA
  • Power dashboards from a shared cube
  • Connect AI agents to governed data via MCP
  • Build OLAP models with Mondrian and Apache Calcite

For BI and analytics teams exposing governed data to humans and AI

Role
mcp-server
Language
Java
Licence
Apache-2.0
Forks
654
Open issues
77
Last push
2026-09-16
Latest release
2.5 · 2013-09-09
topicssemantic-layerolapmdxxmlamcpbusiness-intelligence