Learning level
94/100
Learning level
94/100
Guide type
Security
Audience
Users & builders
Sections
4
Understand the privacy controls used by VauPrime MCP to keep AI-agent responses purpose-limited and avoid unnecessary personal, internal and diagnostic data exposure.
What this guide covers
Quick facts
VauPrime MCP does not intentionally return raw database or RPC payloads. Each tool defines an output schema for the accounting task it performs.
Current minimized outputs remove data that is not required for the tool purpose.
Tool failures are normalized so backend response bodies, SQL/PostgREST/RPC details, tokens and stack traces are not intentionally exposed to connected clients.
The public VauPrime Privacy Policy describes the broader categories of data processed by VauPrime and connected AI/MCP services: https://www.vauprime.com/privacypolicy
Production endpoint
https://mcp.vauprime.com/mcp