DigiQuench

Home / Trust & security

Trust & security

How we protect your data and your decisions

Least-privilege access, encryption, human approval for risky actions, and a full audit trail.

AreaControl
Data accessRead-only exports or database views; only the columns an agent needs; masking at the adapter
HostingAWS, Mumbai region (ap-south-1) for Indian clients; a dedicated AWS environment on request
NetworkOnly HTTPS is open; services listen on localhost behind a reverse proxy; admin access through AWS Systems Manager with no SSH port
EncryptionTLS on every endpoint; S3 and EBS encrypted at rest; credentials encrypted in the orchestration layer
IdentityIAM roles with least privilege; no long-lived AI API keys (Bedrock is called through the server's role); token-protected APIs; read-only public views
ActionsRisk policy with human approval; every decision logged with who and when
AI modelsLanguage models never decide on money, identity or access; no client data in services that train on prompts; India-only model processing available
GuardrailsPersonal-data detection and masking, read-only SQL with blocked write keywords, output length and format checks
MonitoringCloudWatch metrics and alarms, failed-run alerts, spend budgets with email alerts
BackupsNightly copies to S3 with lifecycle rules; restore tested monthly

Roles under India's DPDP Act

You remain the Data Fiduciary for your customers' personal data. DigiQuench acts as a Data Processor on your instructions under a data processing agreement covering purpose, safeguards, breach notification, retention and deletion.

Sub-processors

Amazon Web Services (hosting, storage, email and AI models through Amazon Bedrock), and any other provider you approve in writing. We list them in your agreement and tell you before any change.

Step into DigiQuench.
Agents that heal your data.

Watch an agent find planted problems, fix the safe ones and wait for your approval on the rest. Thirty minutes, live.