Choose your deployment path
StudioLens ships as two builds of one template. They share the transcript-analysis model, but they differ in infrastructure, ingestion, retention and history, the pages that are visible, and whether the PPAC message-credit pages are available. The two builds are not page-for-page identical — pick the one that matches your setup.
The two builds at a glance
| Dataverse (Direct) | Fabric | |
|---|---|---|
| Best for | The simplest, lowest-footprint start — live, recent analysis. | Scale, longer history, and the message-credit pages. |
| Infrastructure | None — parses transcripts live in the Power BI model. | Fabric capacity (F2+ or trial) and a Lakehouse. |
| Ingestion | Live in-model parsing on refresh. | Scheduled Spark ingestion into Delta tables. |
| Transcript pages (quality, topics, flow, errors, feedback, knowledge) | Yes | Yes |
| PPAC message-credit pages | No — the Credit Consumption page ships hidden and is not available to ordinary report users. | Yes — Credit Consumption is a visible page. |
| Data volume / history | Small–medium; bounded by Dataverse’s ~30-day transcript retention. | Large; longer history retained in the Lakehouse. |
| Multi-environment | One or many Dataverse environment URLs. | Via the Lakehouse. |
| Visible page count | 9 visible pages. | 11 visible pages. |
Dataverse (Direct) — simplest
The simplest, lowest-footprint build. Transcripts are parsed live in the Power BI model, so there is no infrastructure to stand up: open the template, set the two required inputs, sign in, and refresh. Best for live, recent analysis within Dataverse’s ~30-day retention window, across one or many Dataverse environments.
Fabric — scale, history, and credits
Scheduled Spark ingestion lands transcripts and message-credit consumption as Delta tables in a Lakehouse. Choose Fabric for larger volumes, longer history beyond the Dataverse retention window, and the PPAC message-credit pages that the Direct build does not expose.