Integration proposal: Virtuosis voice biomarkers for Gemini Live

Virtuosis is an EPFL spin-off focusing on voice biomarkers. We would like to contribute an integration for Gemini Live with this positioning:

> Analyze voice from Gemini Live conversations with Virtuosis to generate voice biomarkers for wellbeing and health insights.

The proposed Python example would:

- capture only the user's 16 kHz PCM16 input stream, never Gemini-generated audio;
- keep the buffer bounded and finalize it as a mono WAV after the live session;
- submit the recording to Virtuosis from the server side, with credentials kept out of the client;
- include tests for audio formatting, duration limits, and an explicit recording-authorization boundary.

We already have the adapter and tests implemented in our private integration repository. Would you prefer this as a new example directory in this repository, or as a documented third-party integration linked from the Partner integrations section?

Related Cookbook proposal: https://github.com/google-gemini/cookbook/issues/1341

Review the proposed Python example requirements and the related Cookbook proposal at issue 1341, then inspect the Partner integrations section to understand the available placement options. The issue is ready only after maintainers choose between a new example directory and a documented third-party integration, with the stated audio, server-side credential, authorization, and test requirements covered.

Written by the indexing model from the issue text.