end userUpdated 2026-08-04

Choose Your Connection

What this covers

There are six ways to get numbers out of Tessallite, and they all read the same governed models. This page helps you pick the right one for the job you are about to do — building an Excel dashboard, writing a Power BI report, exploring in Tableau, or pulling data into a Python notebook — before the walkthrough pages that follow show you each one step by step.

The short version

If you are about to...Use thisWalkthrough
Build or refresh a dashboard inside ExcelTessallite Excel add-in, or a native PivotTable over XMLABuild Your First Excel Dashboard and Build a Report with the Excel Add-in
Drop single governed numbers into Excel cellsTESSALLITE.* worksheet functionsExcel Formulas That Stay Live
Build a shareable report in Power BIPostgreSQL connector with DirectQueryBuild a Power BI Report
Explore visually and build dashboards in TableauPostgreSQL connector, live connectionBuild a Tableau Dashboard
Analyse data in Python or train a model in Jupyterpsycopg2 over the JDBC endpointQuery Tessallite from a Jupyter Notebook
Put governed metrics inside your own appHeadless API (JSON in, rows out)Headless API

Whichever you pick, three things are always true:

  1. The numbers match. Every connection reads the same semantic model, so Excel, Power BI, and a notebook all return the same figure for the same question.
  2. The security follows you. Row security, column restrictions, and your persona apply on every connection. You cannot accidentally see more than you are allowed to.
  3. The speed follows you too. Tessallite routes each query to a pre-computed summary when one exists, on every connection, with no setting to turn on.

The two Excel paths, side by side

Excel gets two entries in the table because they suit different habits. Both are fully supported and both read the same models.

Excel add-inNative PivotTable (XMLA)
What it feels likeA guided task pane inside Excel: drag measures and dimensions, ask questions in plain language, insert resultsThe familiar PivotTable field list connected to a cube
Best forAnalysts who want help assembling a report, or who want the agent inside ExcelAnalysts who already think in PivotTables and want slicers, timelines, and Show Values As
NeedsThe add-in manifest from your administratorA workbook data connection to the XMLA endpoint
Extra tricksTESSALLITE.VALUE cell formulas, KPI inserts, persona switcher, one-click chartsTop 10 value filters, drill-through by double-click, GETPIVOTDATA, CUBE formulas

If you are not sure, start with the add-in: it is the gentler on-ramp, and you can add a native PivotTable later in the same workbook.

Tip. You do not have to choose forever. A workbook can hold an add-in report on one sheet and an XMLA PivotTable on another, and they will agree with each other, because both are views of the same model.

Power BI and Tableau: live, not imported

Both tools connect to Tessallite through the PostgreSQL connector on port 5433. The one setting that matters is the connection mode:

An imported extract still works, but it is a photograph of the data at one moment: it cannot be accelerated by new aggregates, and it drifts out of date until someone refreshes it. Live connections stay current and stay fast.

Jupyter and Python: when a BI tool is not the destination

Sometimes the report is not the end of the work. You might be building features for a machine-learning model, running a statistical check, or feeding a pandas pipeline. For that, Tessallite looks like an ordinary PostgreSQL database on port 5433, and the standard psycopg2 driver connects with your normal Tessallite sign-in. The walkthrough at Query Tessallite from a Jupyter Notebook goes from pip install to a plotted chart and a trained model in one sitting.

The reason to go through Tessallite rather than straight to the source database is the same as everywhere else: the measure definitions, the security rules, and the row filtering are applied before the data reaches your notebook, so the numbers in your experiment reconcile with the numbers on the executive dashboard.

The headless API: for builders, not readers

If you are embedding metrics in your own application — a mobile screen, a customer portal, a scheduled job — the headless API lets your code ask for measures and dimensions by name and get JSON rows back, with the same routing and security as every other path. It is documented for developers at Headless API. If your goal is a chart you look at rather than a product you ship, one of the five paths above is simpler.

What you need for any of them

Trying this on the demo workspace? A local demo install uses workspace slug acme-demo, sign-in admin@acme-demo.com with password acme-demo, and gateway localhost. The walkthroughs in this section use those values and the demo modelx model, so you can follow along click for click.

Related