eMondrian is an open-source semantic layer: a server that holds your business model — cubes, dimensions, hierarchies and measures — and answers questions about it by querying your database.
You describe the model once. Excel, Power BI and AI agents then work with Revenue, Region or On-Time Rate instead of tables, joins and formulas — and all of them get the same numbers.
%%{init: {"flowchart": {"curve": "linear"}}}%% flowchart LR excel["Excel"] -->|XMLA · MDX| em pbi["Power BI"] -->|XMLA · DAX| em ai["AI agents"] -->|MCP| em em["<b>eMondrian</b><br/>semantic model"]:::core -->|SQL| db[("ClickHouse,<br/>PostgreSQL, …")]
How it works
- Relational OLAP (ROLAP). eMondrian doesn’t copy or store your data. Every question becomes SQL against your database, so results are always current.
- Any JDBC database. ClickHouse and PostgreSQL drivers are included; other databases work with their JDBC driver. Column-store databases such as ClickHouse are an especially good fit: ClickHouse does the heavy aggregation, eMondrian provides the business model.
- Built on Mondrian. eMondrian is based on Mondrian, the open-source OLAP engine. The engine was extended to work well with Microsoft clients — for example Excel sessions, calculated members and sets — and with new query languages.
- Caching. Within a client session, repeated queries are answered from eMondrian’s cache without going back to the database.
Three ways to connect
| Client | Protocol | Typical use | Edition |
|---|---|---|---|
| Excel | XMLA / MDX | PivotTables, ad-hoc analysis | Free, open source |
| Power BI | XMLA / DAX | Dashboards and reports over a live connection | DAX module |
| AI agents | MCP | Questions in plain language, schema authoring | MCP module |
Excel and Power BI connect to eMondrian the same way they connect to Microsoft Analysis Services — through the built-in Analysis Services connection. No extra drivers on user machines.
The DAX and MCP modules are included in the Community download and unlocked by a license key — a free 30-day evaluation key is available. See Editions & modules.
When eMondrian is a good fit
- Several reports and tools — Excel, Power BI, AI assistants — must use the same definitions.
- Your data lives in ClickHouse or another SQL database and should stay there: no extracts, no data copies.
- You want an open semantic model in readable XML files that you can review and keep in Git.
- You’re moving away from SSAS Multidimensional or Mondrian / Pentaho and want to keep Excel and Power BI working.
If you have one or two simple Power BI reports on one database, a direct connector may be all you need.
Next steps
- Quickstart — run eMondrian locally and connect Excel, Power BI and an AI agent.
- Concepts — catalogs, cubes, dimensions, hierarchies and measures.