Getting Started
May 24, 2026 ยท View on GitHub
There are two ways to use pbir: directly from the terminal (human), or through an AI coding agent (agent-driven). Both start with installing the CLI.
Install
PyPI (recommended)
uv tool install pbir-cli
Or with pip: pip install pbir-cli
Native installers
Download the latest installer for your platform from GitHub Releases, run it, then open a new terminal.
If your OS blocks the installer
macOS: Try to open the installer once. Go to System Settings => Privacy & Security => scroll to the Security section => click Open Anyway.
Windows: Click More info on the SmartScreen dialog => click Run anyway.
Verify the installation:
pbir --version
Getting Started: Humans
For Power BI users who want to use pbir directly from the terminal.
Terminal basics
- macOS: open
Terminal - Windows: open
PowerShell,Command Prompt, orWindows Terminal cdchanges folders. Quote paths with spaces. PressTabto autocomplete.
Your first 5 minutes
1. Go to the folder containing your PBIR report
cd "C:\Users\YourName\Documents\My Reports"
2. Find reports
pbir ls
3. Inspect the structure
pbir tree "Sales.Report" -v
This shows pages, visuals, and the fields each visual uses.
4. Inspect the semantic model
pbir model "Sales.Report" -d
Use this before adding or rebinding visuals so you know the table, column, and measure names available.
5. Make a change
pbir add visual card "Sales.Report/Overview.Page" --title "Revenue" -d "Values:Sales.Revenue"
6. Validate
pbir validate "Sales.Report"
Path syntax
pbir addresses report objects with filesystem-style paths:
Report.Report/Page.Page/Visual.Visual
Examples:
Sales.Report # A report
Sales.Report/Overview.Page # A page
Sales.Report/Overview.Page/Revenue.Visual # A visual
Sales.Report/**/*.Visual # All visuals (glob)
Quote paths when names contain spaces.
Safe workflow
pbir backup "Sales.Report" # Before edits
# ... make changes ...
pbir validate "Sales.Report" # After edits
pbir restore "Sales.Report" # If something goes wrong
Getting Started: Agents
For users of Claude Code, GitHub Copilot, Cursor, Gemini CLI, Codex, or similar AI coding agents.
Step 1: Install skills
Your agent needs skills that teach it how to use pbir effectively. Install the full power-bi-agentic-development marketplace, which includes skills for reports, semantic models, themes, DAX, Fabric CLI, and more.
# Claude Code
claude plugin marketplace add data-goblin/power-bi-agentic-development
# GitHub Copilot CLI
copilot plugin install data-goblin/power-bi-agentic-development
Step 2: Navigate to your report
Open your agent in the folder containing your PBIR report. The agent needs to be in the right directory to find reports.
cd "path/to/my/reports"
Step 3: Ask your agent
Once skills are installed, just describe what you want. The agent will use pbir commands under the hood. Examples:
- "List all reports and show me the structure of Sales.Report"
- "Add a card visual showing Revenue to the Overview page"
- "Set all title font sizes to 14 across the report"
- "Create a new report connected to spaceparts-dev/spaceparts-otc-full.SemanticModel"
- "Add a TopN filter for the top 10 customers by revenue"
- "Validate the report and fix any issues"
The agent will discover properties, bind fields, format visuals, and validate -- all through the CLI. You don't need to know the commands yourself.
Step 4: Review and validate
After the agent makes changes, always ask it to validate:
- "Validate the report"
- "Show me the tree view with fields"
Or run it yourself:
pbir validate "Sales.Report" --all
Next Steps
- CLI Docs -- full command reference
- CLI Workflows -- end-to-end workflows
- power-bi-agentic-development -- full plugin marketplace
Help
pbir --helppbir <command> --help- Report issues