Get started with Weaver

Works natively with Microsoft Fabric. Nothing else to host.

Install Weaver with pip in a Fabric Notebook or your own Python environment. Use your existing Fabric workspaces, compute and permissions, without a separate Weaver service, account or software licence.

Install the package

Weaver supports Python 3.13, as well as Python 3.11 and 3.12.

Install Weaver
python -m pip install --upgrade weaverstack
weaver --version

Check the installed version before running your first project.

Existing Fabric compute and workspace permissions still apply.

Installation details →

Work in Fabric or your own editor

Desktop Python

Use your terminal, VS Code or any Python environment on Windows, macOS or Linux. Write locally and run Build, Load, Test and Health against Microsoft Fabric.

Native Power BI semantic models and Reports deploy without Power BI Desktop or a Windows computer. Weaver deploys Report definitions; it is not a visual Report designer.

Fabric Notebook

Install the same Python package:

Notebook cell
%pip install weaverstack

Then:

Python
import weaver

Weaver can use the Notebook's existing workspace, Spark session and identity. No desktop installation is required.

For repeatable jobs, add Weaver to the attached Fabric Environment rather than relying on a session-scoped installation.

Both

Develop locally, run in Fabric, or move between the two. Your Weaver project doesn't need to be rewritten for a different execution environment.

Working between desktop and Fabric →

Sign in to Fabric

Browser sign-in
For interactive development.
Azure CLI
When you're already authenticated with Azure.
Service principals
For unattended jobs and automation.
Fabric Notebook identity
When running inside your workspace.

No separate Weaver account or authentication service is required. Existing Fabric and Power BI permissions still apply.

Start with what you're working on

A Lakehouse table, a Warehouse query or a Power BI project is enough to begin. Pick one to try Weaver.

Try the bundled Sales example

Use an existing disposable workspace in place of Analytics.

Terminal
weaver initialise --workspace Analytics --project-folder ./sales \
  --warehouse Curated --example Sales --non-interactive
cd sales
weaver workflow full

Lakehouse

Use Python or Spark SQL to load files into Lakehouse tables.

Build your first Lakehouse pipeline →

Warehouse

Write a T-SQL query. Weaver creates the table and the loading steps.

Build your first Warehouse pipeline →

Semantic Models

Deploy standard Power BI models, apply shared rules and unit test calculations.

Deploy your first Semantic Model →

Reports

Deploy the same Power BI Report to different workspaces, connected to the right model.

Deploy a Power BI Report →

Start in a Power BI service workspace

With an existing Power BI service workspace and the appropriate licence and permissions, you can deploy native Semantic Models and Reports, apply modelling conventions and validate DAX calculations. No Fabric Lakehouse or Warehouse workloads required.

Start with a Power BI model or Report. You do not need to build a Lakehouse or Warehouse first.

Get started with your existing Power BI project →