This page is maintained by KleinHub AI to answer common security and privacy questions about our product. It reflects our current practices and is not an independent certification.
When you use the microphone or enable voice replies in the AI Copilot, your audio is sent to our servers and forwarded to the Lovable AI Gateway, which routes it to third-party speech-to-text and text-to-speech providers (currently OpenAI models) solely to transcribe your request and synthesize the reply. We do not retain the audio; only the resulting transcript is kept with your chat history. Voice is optional — text chat remains fully functional if you decline microphone permission, disable voice replies, or run out of credits.
KleinHub AI operates two independent oversight layers, kept separate by design.
Every workspace has a built-in Guardian that reviews AI actions before they run, flags unusual spending or agent behaviour, and gives the workspace owner an audit trail of every automated action taken on their behalf. High-severity events can auto-pause the workforce until a human reviews.
A separate platform-level Guardian, staffed by authorised KleinHub personnel, watches for patterns that cross accounts — such as abuse, provider outages, or repeat compliance signals. Every access of workspace data by platform staff is recorded in an internal admin access log.
Guardian findings are risk signals, not legal conclusions. This page is maintained by the KleinHub AI team and is not a certification.
Request access, export, deletion, or correction of your personal data. We aim to respond within 30 days.