Privacy Policy
Privacy-first demo: the microphone is started explicitly by you, visibly indicated while active, automatically stops after inactivity, and is stopped when the page loses focus. It does not automatically resume.
1. What this policy covers
This policy explains how the Corporate Agent Network demo handles information when you use the website, architecture demonstration and Live Voice features.
2. Information you provide
You may provide spoken words, recognised transcripts, typed text, language and style selections, and other inputs used to demonstrate the agent workflows. Please do not submit confidential or highly sensitive information to this public demo.
3. Microphone and speech recognition
The site requests microphone access only when you start the microphone. Speech recognition is provided through browser speech-recognition capabilities. Depending on your browser and device, audio may be processed by the browser vendor or its speech-recognition service under that provider's own privacy terms.
The demo is designed to stop microphone capture after approximately 20 seconds of silence and when the page is hidden, the browser window loses focus, or you leave the page. Returning to the page does not automatically restart the microphone.
4. AI processing
Recognised text and other submitted text may be sent to AI services as necessary to perform translation, editing, agent coordination or related demo functions. Voice scripts sent to the spoken-output feature are processed by OpenAI to generate AI speech.
The demo also uses Mozaik runtime/infrastructure components for multi-agent execution and observability, and Render for hosting. These providers may process technical and request data required to provide their services, subject to their own terms and privacy practices.
5. Storage and retention in this demo
The Live Voice demo does not use an application database to permanently store your transcripts. Working state and generated-audio cache are held transiently in server memory and can disappear when the service restarts. Operational logs may contain technical metadata needed to run and diagnose the service.
Third-party providers may independently retain or process data according to their own policies. This project does not control those provider-side retention periods.
6. Cookies and tracking
The demo does not intentionally add advertising cookies or behavioural advertising trackers. Hosting, browser or infrastructure providers may still process standard connection information such as IP address, request time, user agent and diagnostic data.
7. Why information is processed
Information is processed to provide the feature you request, demonstrate the multi-agent architecture, generate translations and spoken output, keep the service secure, and diagnose technical problems.
8. Sharing
Information is shared only as needed with service providers that operate the demo's hosting, browser speech recognition, agent runtime and AI functionality, or when disclosure is required by law. The project does not sell user data.
9. Your choices
You can use the architecture demo without enabling the microphone. You can stop microphone access at any time, disable voice output, clear the current voice session, revoke microphone permission in your browser, or stop using the service.
10. Security and limitations
Reasonable technical measures are used for this prototype, but no public internet demo can guarantee absolute security. Do not use the service as a repository for confidential corporate, personal or regulated data.
11. Children
This service is intended as a professional technology demonstration and is not directed to children.
12. Changes
This policy may be updated as the prototype evolves. The effective date above identifies the current published version.
13. Contact
Project owner: Oleksandr Koretskiy, Koretskiy Consulting, Serbia. Privacy questions can be sent to office@koretskiy.rs. Project website: koretskiy.com.