Privacy Policy & Data Sovereignty Charter

1. Information We Collect & Process

When you create an account and operate an Adhyakṣa workspace, we process the following categories of information:

• Account & Organization Identity: Master owner name, work email address, company title, organization workspace slug, and billing details.
• Operational Enterprise Data: Invoices, inventory SKU catalogs, cashier receipts, employee rosters, payroll deductions, debt facility schedules, and CRM notes stored in your tenant PostgreSQL database.
• Document Studio Files: Bank statements (PDF, CSV, Excel) uploaded for optical character recognition (OCR) and transaction reconciliation.

2. Processing Purposes & Zero Monetization Guarantee

HostingSpell processes Customer Data exclusively to deliver the functionality of the Adhyakṣa Unified Management OS.

3. Private AI Direct Routing & Zero-Retention Commitment

Adhyakṣa utilizes a sovereign direct-to-provider AI architecture. When you configure API keys for OpenRouter, Anthropic Claude, OpenAI, Replicate, or self-hosted Ollama:

• Direct Point-to-Point Routing: Prompt payloads stream directly from your backend instance to your designated LLM provider endpoint.
• Zero AI Training: Your business records, financial statements, and staff queries are NEVER used to train public foundation models.
• Zero Token Storage: We do not log or store prompt completions on central intermediary proxies.

4. Multi-Tenant Scoped Cryptographic Isolation

Every record in Adhyakṣa is bound to an immutable system_id tenant scope. PostgreSQL Row-Level Security (RLS) policies enforce mathematical data isolation at the database kernel level, preventing any cross-tenant data leakage between neighboring organizations.

5. Continuous Snapshots & 1-Click Rollback Storage

To support 1-Click Atomic Rollbacks, Adhyakṣa records encrypted entity diffs in the action_execution_log table. These rollback snapshots are stored in encrypted form and purged automatically upon the expiration of your tier's retention window (24 hours on Pranava, 30 days on Sthira, 1 year on Ananta).

6. Perpetual Right to Instant PostgreSQL Export & Total Deletion

In adherence to our Data Sovereignty Charter, you have the perpetual right to export your complete database tables in open SQL, CSV, or JSON formats directly from the system settings console. If you terminate your account, all tenant records are permanently expunged from production SSD storage within thirty (30) days.

7. Biometric & Medical Practice Confidentiality

For organizations utilizing our HRM Biometrics and Practice Management engines:

• • Biometric Hashes: Fingerprint and RFID clock-in records are stored solely as one-way cryptographic verification hashes, not raw biometric images.
• • Clinical Confidentiality: Doctor-patient consultation notes, clinical attachments, and attorney matter files are shielded with row-level permission tokens and encrypted at rest with AES-256.

8. Infrastructure, Subprocessors & Dokploy Hosting

Cloud instances are hosted on ISO 27001 and SOC 2 Type II certified datacenters operated by HostingSpell. For enterprise customers demanding zero cloud exposure, Adhyakṣa may be deployed self-hosted on your own private bare-metal servers or air-gapped Dokploy clusters with zero external dependencies.

9. Global Privacy Frameworks (DPDP Act 2023, GDPR & CCPA)

Adhyakṣa is engineered to comply with international privacy legislation:

10. Security Protocols, Encryption & WORM Logs

All data transmission between your browser and Adhyakṣa servers is encrypted using TLS 1.3 with Perfect Forward Secrecy (PFS). All database volumes and document attachments are encrypted at rest using AES-256. Sovereign enterprise tiers include Write-Once-Read-Many (WORM) immutable audit logs for regulatory compliance.

11. Data Protection Officer (DPO) Inquiries & Contact

For questions regarding this Privacy Policy, to exercise your data subject rights, or to request an enterprise Data Processing Agreement (DPA), please contact our Data Protection Officer: