Domain controls
Public AI service embeds can be restricted to approved hostnames, reducing unintended usage outside the business domain.
Allowed domains
Embed host checks
Public route guardrails
Last updated June 2026
TL;DRSarvadrik AI includes domain allowlists, owner-scoped records, server-side API keys, support logs, and human handoff workflows. These controls help businesses deploy AI services with practical guardrails.
Security
Controls for safer AI service deployment.
Security in a business AI service starts with scoped access, approved knowledge, domain controls, server-side keys, and logs that make failures visible.
SAT
Sarvadrik AI Team
AI Services Platform· Updated June 2026
Access
User-scoped records
Embeds
Allowed domains
Logs
Latency and errors
Handoffs
Traceable leads
Scoped data
Services, conversations, messages, leads, and logs are tied to owners so dashboard views can load account-specific data.
Owner records
Service-level filtering
Conversation linking
Operational logs
Completions, fallback behavior, webhook outcomes, latency, and status codes can be reviewed for production troubleshooting.
Support logs
Webhook status
Provider failures
Human handoff
The product supports escalation and intake workflows instead of forcing AI to answer every request.
Contact intent
Lead records
Resolved sessions
Production readiness checklist
Set allowed domains for every public AI service embed.
Store service-role keys only in server environments.
Confirm OpenRouter and Supabase environment variables are present.
Run the E2E script against production after deployment.
Security posture
Sarvadrik AI focuses on practical controls: restricted embeds, account-scoped records, logged failures, and clear operator review paths.
Domain allowlists
Server-side keys
Logged failures
Scoped dashboard data