Implemented safeguards
The OpenAI API key is used server-side, input is validated before generation, and production-facing errors avoid exposing provider details or stack traces.
Security
ReplyWise is currently a focused MVP. This page documents implemented safeguards and known production gaps.
Data storage
ReplyWise stores authenticated account details, workspace records, settings, usage events, and audit records in Supabase. The app does not use a committed storage upload runtime or an external analytics platform.
AI safety
The advisor prompt and internal product logic are designed to avoid aggressive, deceptive, manipulative, illegal, or unsafe business replies.
Security contact
A public vulnerability-reporting channel is not configured. Do not send sensitive data through an unverified contact channel.
Operational boundaries
Formal incident response and SOC2 controls are not represented as implemented in this repository.