Thanks — you're asking which Speech-to-Text (STT) models are hosted inside a Virtual Private Cloud (VPC). I looked into this for you. There is no documented information that any specific STT models are hosted within a VPC. LiveKit’s infrastructure and some services run in VPCs (for example on GCP, Akamai, and Digital Ocean), and there are regional STT deployments (for example Deepgram models co-located in Mumbai), but none of the sources state that particular STT models are hosted inside a VPC. If you need a firm answer for compliance or network design, I can help with next steps: • Do you need confirmation for a specific model/provider (for example, Deepgram) or for all STT models used in your account? • Would a written confirmation for audit purposes help — and if so, do you need it scoped to a region or project? Tell me which of those you'd like and I’ll proceed. Sources: Security Overview | LiveKit | Building Performant Voice Agents in India | LiveKit