pbi-guard
github.com/patens-dev/pbi-guard- Category
- Developer Tools
- Rank
- No. 4185Tools index
Previous survey · No. 3712 ·
- Pricing
- Open Source
- Type
- TOOL
- Use case
- Software Testing & Quality
- GitHub
- 1 stars
- Latest release
- v0.1.5
- Date
About
pbi-guard is an open-source, MIT-licensed testing and governance tool for Power BI semantic models (.pbip/TMDL projects), acting like pytest for Power BI. It runs DAX logic checks, tabular-modeling assertions, and produces auditable UAT sign-off reports locally in milliseconds, catching bad DAX, silent measure regressions, and human errors (such as arbitrary scalar multipliers) before a pull request reaches production, without needing Power BI Desktop, complex regular expressions, or cloud credentials.
At a glance
PBI Guard checks Power BI semantic-model definitions against rules you set. It is worth evaluating if several people edit the same measures and you want to catch a formula change before it becomes a reporting problem.
Good fit
Teams that keep Power BI projects in the file-based PBIP format and can define the formulas or conventions that matter.
Consider something else when
Use refresh and reconciliation tests when the question is whether the model produces the right numbers on real data.
Start here
Use a disposable project with one protected measure. Change its formula and check whether the configured rule catches the edit on save.
Cost and requirements
The project documents Power BI Desktop integration on Windows and Python 3.10 or newer for developer setup.
Limits to know
It inspects model definitions and rules without running Analysis Services. A passing check does not verify a refresh, an aggregation or the business logic behind the rule.
Sources checked 2026-11-05. Source-based assessment.
Why it made the leaderboard
Brings pytest-style contract testing to Power BI semantic models, catching bad DAX and silent measure regressions in TMDL files before they reach production dashboards, a gap with no comparable existing tool.
Tags
Tech Stack
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.