Security & Trust Center

Built for enterprise trust.
Verified by architecture.

StreetMP OS governance commitments are enforced at the infrastructure level — not by policy configuration. This page documents how we protect your organization's AI operations.

0 bytes
Data retained after scan
Zero-retention by architecture — not policy
100%
Audit events HMAC-signed
Cryptographic integrity on every record
AES-256
Encryption at rest
All tenant data encrypted, key-per-tenant
Audit log retention
Append-only, tamper-evident, never purged

Security Architecture

How we protect your data at every layer of the stack.

Zero-Retention Architecture

StreetMP OS never stores the content of AI prompts. The governance engine intercepts, scans, and sanitizes data in memory. Only cryptographic audit receipts (SHA-256 hash + verdict + entity metadata) are persisted to the audit ledger — never the raw text. This is an architectural guarantee, not a configurable policy.

Encryption Standards

All data at rest is encrypted using AES-256-GCM with per-tenant encryption keys managed via a dedicated key service. Data in transit uses TLS 1.3 exclusively. The NeMo CLAW scanning engine operates on ephemeral memory — no disk writes during the scan pipeline. Encryption keys are rotated on a 90-day schedule.

Data Residency & APAC Sovereignty

StreetMP OS supports regional data residency for India (ap-south-1 Mumbai), Singapore (ap-southeast-1), and Malaysia (ap-southeast-3). No personally identifiable data crosses regional boundaries. APAC enterprise customers can enforce strict data localization — all governance processing occurs within the customer's selected region.

Browser Extension Privacy

The Sovereign Shield browser extension operates as a transparent proxy — it intercepts AI submissions before they reach the provider, applies governance rules, and forwards the sanitized version. The extension does not log browsing history, page content, or any data outside the AI submission flow. Extension source is available for security review under NDA.

Access Control & Multi-Tenancy

Every API call, dashboard request, and governance action is tenant-isolated at the database level using Row Level Security. There is no shared query path between tenants. Role-Based Access Control enforces OWNER / ADMIN / ANALYST / AUDITOR / MEMBER permissions with cryptographically signed session tokens. MFA is enforced for all admin roles.

Audit Trail Integrity

Every governance event — scan verdict, policy change, kill switch toggle, compliance export — is written to an append-only audit ledger with HMAC-SHA256 chaining. The chain ensures retroactive tampering is detectable. Audit logs can be exported in JSON or CSV format, signed with a timestamp authority certificate, for use in external audits.

Compliance Frameworks

Regulatory alignment across global and APAC markets.

SOC 2 Type II
Global

Security, Availability, Confidentiality controls documented and evidence-exportable.

Audit-Ready
ISO 27001:2022
Global

Information Security Management System alignment with controls mapping available.

In Progress
GDPR
EU

Lawful basis documentation, data subject rights, DPA templates available on request.

Audit-Ready
DPDP 2023
India

India Digital Personal Data Protection Act — zero cross-border transfer by default.

Certified
MAS TRM
Singapore

Monetary Authority of Singapore Technology Risk Management alignment complete.

Audit-Ready
PDPA
Malaysia/SG

Full PDPA (Malaysia) and PDPA (Singapore) alignment with data localization options.

Audit-Ready
HIPAA
US

Healthcare data safeguard controls — BAA available for healthcare enterprise customers.

In Progress

No certifications are claimed until formally issued. Status reflects current control completion and audit readiness.

Enterprise Deployment Models

Choose the deployment architecture that matches your security requirements.

Cloud SaaS99.9% uptime SLA

Fully managed. Zero infrastructure. SOC2-compliant shared environment with strict tenant isolation.

Private CloudCustom SLA

Dedicated compute in your cloud account (AWS / GCP / Azure). All governance data stays in your VPC.

On-PremiseSelf-managed

Full self-hosted deployment on your infrastructure. Air-gapped environments supported. Suitable for regulated industries.

HybridNegotiated SLA

Scanning engine on-premise, audit ledger and dashboard cloud-hosted. Common in BFSI and healthcare deployments.

Responsible AI Commitments

We do not train AI models on customer data — ever.
We publish detection model accuracy metrics quarterly.
False positive reporting is surfaced to governance teams within 24 hours.
All policy changes require admin authentication and are audit-logged.
Kill switches allow instant suspension of AI access for any user group.
We report to customers within 72 hours of any governance-relevant incident.
We support customer-controlled data deletion on contract termination.
No behavioral profiling of individual employees outside governance scope.

Vendor Security FAQ

Answers to common questions from enterprise security teams.

Ready for your security review?

Request a technical security briefing, architecture walkthrough, or compliance evidence package.