Self-Hosted BI Pixie User Guide
BI Pixie measures how your audience uses your Power BI reports, from page views down to the visuals, filters and slicers people interact with. BI Pixie scans each report, page and visual for dynamic and static hyperlinks, filters, slicers, bookmarks, tooltips, and drill-through configurations. It then adds Pixies to your reports and semantic models: tiny, native Power BI visuals and hidden measures through which BI Pixie records what people do in the report. Pixies are hidden from the people who view your reports.
You can add Pixies to as many reports as you need, and run BI Pixie again on the same reports whenever they change, so it fits into your BI application lifecycle management or CI/CD process. Once you publish the reports, BI Pixie collects usage and engagement events from every Power BI client your audience uses: workspaces, apps, embedded reports, mobile, and desktop.
BI Pixie collects viewer user names only when include_user is true in bi_pixie_config.json, and it then keeps the email address, or a masked ID when MASK_USER_NAME is true on the function app. When include_user is false, it counts viewers by a masked IP address.
This user guide covers self-hosted BI Pixie from setup to engagement analysis.
AI Readiness is in BI Pixie Cloud and the Workload. Scoring your semantic models for Copilot and Fabric data agents, the optimizations BI Pixie proposes, and benchmarks are part of BI Pixie Cloud and the BI Pixie Workload for Microsoft Fabric. A self-hosted deployment collects the usage telemetry, and your AI agents can query it through the BI Pixie MCP Server.
BI Pixie components
BI Pixie offers two serverless deployments that keep your audience's engagement data inside your organization: the Power Platform deployment, also known as the Starter version, and the Azure deployment.
Both deployments need BI Pixie Desktop, marked as item 1 in the diagram above. The tool runs on a local machine and adds Pixies to your Power BI project files, so that BI Pixie can track usage and engagement once you publish the reports.
BI Pixie Desktop is the new name of BI Pixie Instrumentation. The file is still named bi-pixie.exe. Version 3.1.2 and earlier end a run with the log line “BI Pixie Instrumentation is completed”. Later versions end it with “BI Pixie Desktop has finished”. The default commit message is still “BI Pixie instrumentation”.
You deploy the Power Platform (Starter) version in your Power Platform environment. It suits a small number of reports that serve a mid-size BI audience, and it includes three components:
- Power Automate web trigger cloud flows that collect the engagement events from your BI audience, marked as item 2 on the left side of the diagram above.
- Dataverse tables that store the events and inventory, item 3 on the left side of the diagram above.
- A Power BI app that connects to the Dataverse tables and analyzes the usage and engagement of your BI audience, item 4 on the left side of the diagram above.
You deploy the Azure version in your Azure tenant, and it scales to the high usage of large BI audiences. It includes the following components:
- One or two Azure Function Apps that collect the events, marked as item 2 on the right side of the diagram above.
- A Data Lake Storage Gen2 account that stores your BI audience events, item 3 on the right side of the diagram above.
- A Power BI app that connects to the Data Lake storage and analyzes the usage and engagement of your BI audience, item 4 on the right side of the diagram above.
The Azure deployment carries the full set of features. The following table summarizes the key capabilities of each deployment:
| Features (Source) | Power Platform Deployment | Azure Deployment |
|---|---|---|
| Usage events (Pixies) | Yes | Yes |
| Feedback Controls (Pixies) | Yes | Yes |
| Embedded Surveys (Pixies) | Yes | Yes |
| Data Auditing (Pixies) | Yes | Yes |
| Performance (Azure Log Analytics) See Analyze Performance. | No | Yes |
| Producers and Consumers (Power BI Activity Logs) | No | Yes |
| Business Outcome by Endorsement and Data Sources (Power BI Admin Scan API) | No | Yes |
| Export Auditing (Pixies + Azure Log Analytics) | No | Yes |
| RLS Auditing (Pixies) | No | Yes |
Set up BI Pixie
Before you set up BI Pixie, we recommend installing the BI Pixie Dashboard in your Power BI account and exploring it with sample data. The BI Pixie Dashboard is a Power BI app on Microsoft AppSource, in two versions: Starter and Azure.
To set up BI Pixie, follow these steps:
- Deploy BI Pixie to collect usage and engagement events.
- For the Starter version, see Deploy BI Pixie on Power Platform.
- For the Azure version, see Deploy BI Pixie on Azure.
- Add Pixies to your Power BI reports, as described in Add Pixies to your reports.
- Test the reports and publish them, as described in Test your reports.
- Connect the BI Pixie Dashboard, as described in Set up BI Pixie Dashboard.
- Learn about the visuals in the BI Pixie Dashboard in A Walkthrough of BI Pixie Dashboard.