Security

Security & Trust

Enterprise-grade security for your restaurant data

Last Updated: March 2, 2026

AES-256 Encryption
TLS 1.2+ Transit
SOC2-Ready
GDPR Compliant

At Nirvah AI, security isn't an afterthought—it's foundational to everything we build. We understand that you're trusting us with your business data, and we take that responsibility seriously.

Nirvah AI's use and transfer of information received from Google APIs adheres to the Google API Services User Data Policy, including the Limited Use requirements. See our Privacy Policy for full details.

Encryption at Rest

AES-256

All sensitive credentials (OAuth tokens, webhook secrets) and customer PII (phone numbers, names, email addresses, delivery addresses) are encrypted using AES-256-GCM at the application level before storage. Encryption keys are managed through secure environment variables and never stored in the database.

  • Database encryption: All customer data in our PostgreSQL databases is encrypted at the storage level
  • Application-level encryption: OAuth tokens and customer PII are individually encrypted using AES-256-GCM before being written to the database
  • File storage: Uploaded documents, menus, and knowledge base files are encrypted before storage
  • Backups: All database backups are encrypted using the same standards

Encryption in Transit

TLS 1.2+

All data transmitted between your browser and our servers is protected using TLS 1.2 or higher encryption protocols.

  • HTTPS everywhere: All connections to our platform are encrypted with TLS certificates
  • API encryption: All API calls to Google Business Profile, KitchenHub, and Square are encrypted using TLS 1.2+
  • Voice data: Real-time voice streams are encrypted end-to-end through our voice AI provider
  • Certificate management: We use automated certificate rotation and monitoring

Multi-Tenant Data Isolation

Row Level Security

Nirvah AI employs strict logical separation of customer data. Restaurant data from Google Business Profile, Square POS, and KitchenHub is isolated at the database level using Row Level Security (RLS). Cross-tenant data access is architecturally impossible.

  • Row Level Security (RLS): Database-level policies ensure that queries can only access data belonging to your organization
  • Tenant isolation:Each customer's data is logically separated with unique identifiers tied to the restaurant entity
  • Compliance logging: Compliance actions are logged with full audit trails
  • Periodic reviews: We conduct periodic security reviews and dependency updates

SOC2 Readiness: Our infrastructure and practices are designed to meet SOC2 Type II requirements. We are actively working toward formal certification.

Google Data Deletion: Google Business Profile review data is immediately purged from all systems when the integration is disconnected, in compliance with the Google API Services User Data Policy.

Access Controls & Authentication

We implement multiple layers of access control to protect your account:

  • Secure authentication: Powered by Supabase Auth with support for email/password and Google OAuth
  • Session management: httpOnly, secure cookies with SameSite protection against cross-site attacks
  • Role-based access: Different permission levels for owners, managers, and staff
  • Admin verification: Additional verification codes required for administrative access
  • Token encryption: Integration tokens are encrypted at the application level using AES-256-GCM before database storage

AI & Data Processing Guarantees

We enforce strict boundaries on how customer data is processed by AI systems:

  • Receipt OCR: Receipt images processed via OpenAI Vision are used for transient OCR only and are not used to train global models
  • Voice transcription:Voice transcriptions via Deepgram/LiveKit are encrypted and stored only for the duration required by the restaurant's retention policy
  • Zero-training guarantee:No customer data is used to train third-party AI providers' public foundation models. All AI inference uses enterprise API access with zero-retention terms

Vendor Security

We carefully vet all third-party vendors and require them to meet our security standards:

VendorPurposeCompliance
SupabaseDatabase & AuthSOC2 Type II
VercelHosting & EdgeSOC2 Type II
OpenAI / AnthropicAI Inference & Receipt OCR (Vision)SOC2 Type II
LiveKitVoice AI InfrastructureSOC2 Type II
DeepgramSpeech-to-TextSOC2 Type II
CartesiaText-to-SpeechSOC2 Type II
StripePaymentsPCI DSS Level 1
Google Cloud PlatformAPI Infrastructure & Business ProfileSOC2 Type II, ISO 27001
SquarePOS Data IntegrationPCI DSS Level 1, ISO 27001
KitchenHubPOS Middleware (Toast & Clover)Enterprise Security
TwilioSMS & Voice TelephonyISO 27001

Incident Response

We maintain a comprehensive incident response plan to quickly address any security events:

  • 24/7 monitoring: Automated alerting for suspicious activities and system anomalies
  • Rapid response: Defined procedures for containment, investigation, and remediation
  • Customer notification: Timely communication if your data is affected by any security incident
  • Post-incident review: Root cause analysis and improvements after every incident

Compliance & Certifications

We design our systems to meet or exceed industry compliance requirements:

GDPR

Compliant

Data protection rights for EU residents

CCPA

Compliant

Privacy rights for California residents

SOC2 Type II

In Progress

Security, availability, and confidentiality

TCPA

Compliant

Telephone consumer protection

Data Retention Policy

We retain data only as long as necessary for the services we provide. Below are our standard retention periods:

Data TypeRetention Period
Call recordings90 days
Call transcripts1 year
Order data2 years
Google Business Profile dataUntil integration disconnect or account deletion. All synced review data is immediately deleted upon disconnection.
Restaurant leadsUntil deleted by user
Log data30 days

Security Contact

If you discover a security vulnerability or have security-related questions, please contact us immediately:

Security Team

Email: help@nirvah.ai

For general support: help@nirvah.ai

We take all security reports seriously and will respond within 24 hours.