Security Overview
Last updated: August 23, 2025
Our Security Principles
- Security by design: Minimal access, least privilege, auditability, and defense in depth
- Privacy first: Only collect data required to provide the service; configurable data retention
- Transparency: Clear data flows, storage locations, and subprocessors
Data Protection
- Encryption in transit: All traffic over TLS 1.2+
- Encryption at rest: Databases and object storage encrypted using provider-managed keys (AES‑256 or equivalent)
- Secret management: Application secrets stored in environment variables and rotated per environment
- Data residency & backups: Backups are encrypted and retained per policy; configurable retention windows
Application Security
- Authentication: Token-based auth (Sanctum) with optional multi-factor verification (phone/SMS OTP)
- Authorization: Role-based access controls and plan-based feature gating
- Rate limiting & abuse protection: Tiered rate limits per plan to protect from brute force and abuse
- Input validation: Strict validation on all endpoints; server-side allowlisting for uploads
- Logging & monitoring: Structured logs, error tracking, and anomaly alerts for security‑sensitive operations
AI & Vector Security
- Isolation: Vector namespaces are isolated per user/organization
- Content handling: Uploaded documents are processed in ephemeral workers; only extracted text and metadata required for RAG are stored
- Provider security: Third‑party LLM/embedding providers are accessed strictly over TLS with API keys scoped per environment
Data Privacy Commitment:
We do not use your documents to train our AI models. Your content remains private and is used solely to provide you with analysis services.
Operational Security
- Least privilege: Separate IAM roles for app, storage, and backups; production access is restricted and logged
- Change management: Version control and CI/CD with required reviews for production changes
- Vulnerability management: Regular dependency updates and security patches; periodic security reviews
Compliance & Legal
ArX AI follows industry best practices and is designed to be compliance‑ready. We can support customer assessments and data processing agreements upon request.
- GDPR‑ready: Data subject rights tooling; data export/deletion on request; subprocessor transparency
- Pakistan regulations: Aligned with Prevention of Electronic Crimes Act (PECA) 2016, State Bank of Pakistan (SBP) cybersecurity guidelines, and Personal Data Protection Bill principles for collection, purpose limitation, and security safeguards
- Security controls: Access control, encryption, audit logs, and incident procedures aligned with common frameworks (ISO 27001 alignment)
Customer Controls
- Configurable token limits, rate limits, and phone-verification requirements
- Per‑user token/session management and forced revocation from the dashboard
- Project/document scoping for searches; organization‑level namespaces for business plans
Incident Response
- Detection: Continuous monitoring and alerting for anomalies
- Containment and recovery: Defined runbooks for isolating affected components, rotating credentials, and restoring from clean backups
- Notification: We notify affected customers in a timely manner when legally required
Contact Information
For security inquiries, vulnerability disclosures, or data processing agreement requests, please contact us:
Ali Raza Technologies (SMC-Pvt) Limited
- Security Team:
- General Inquiries:
- Support:
- Phone:
- Address: Chak No 108 G.B, Tehsil Jaranwala, District Faisalabad, Punjab, Pakistan, 38000
- Registration: Registered under Companies Act 2017
Security Disclosure: If you discover a security vulnerability, please report it responsibly to
This document describes our current security posture and may evolve without prior notice. For contractual commitments, refer to your service agreement or data processing agreement.
Additional Resources: