Legal

PRIVACY
POLICY

Last updated: July 2, 2026

1. Who We Are

RGX Systems is a wholesale B2B utility infrastructure provider. We provide API access — ingest, processing, and usage tracking endpoints — to approved Value-Added Resellers (VARs). VARs build their own products and services on top of our infrastructure and are solely responsible for their relationships with their own clients.

This policy covers data RGX Systems collects about its VAR operators. It does not govern the data VARs process on behalf of their own clients — VARs are independently responsible for their data practices and compliance obligations toward their end users.

2. Data We Collect

VAR Account Information

• Company name, contact name, and business email address provided at application
• Billing information processed securely by Stripe (we never store card details)
• VAR node configuration: plan tier, base fee, seat fee, and notes
• API key metadata: key prefix, creation date, last-used timestamp, and status (active/revoked) — we never store the plaintext key

API Usage Data

Every API request authenticated against your VAR node is logged for billing and operational purposes:

• Endpoint called and HTTP method
• Timestamp and response status code
• Payload size in bytes and message count
• Processing units consumed and estimated processing cost (when pipeline processing is enabled)
• Processing time in milliseconds
• Source IP address of the API caller
• Metadata fields you pass in the request body (channel, reference IDs, etc.)

We do not log the content of your payloads — message text, contact names, and other payload data are processed in memory and not persisted to our database unless you explicitly pass them as metadata fields.

Application Data

• Information submitted in your VAR access application (business description, client base, pricing plan, etc.)
• Communications with our team relating to your account

3. How We Use Your Data

• To provision and manage your VAR node and API key
• To calculate monthly billing: base fee plus $20 per active client seat
• To generate usage summaries and billing reports accessible via the /usage endpoint
• To authenticate API requests by verifying your key hash
• To detect misuse, abuse, or security incidents on your node
• To send transactional emails: provisioning confirmations, key rotation notices, and invoices
• To diagnose technical issues and maintain infrastructure reliability

We do not sell your data. We do not use payload data for model training or third-party profiling. Your business data is yours.

4. PHI Handling and HIPAA Compliance

How PHI Is Processed

When a VAR submits a request to /api/v1/process with config.industry: "healthcare":

• The input is received over TLS 1.3 and held in process memory only
• Our scrubPHI() function runs entirely in RAM, identifying and replacing HIPAA identifiers (SSNs, DOBs, MRNs, Insurance IDs, NPIs, phone numbers, email addresses, and street addresses) with bracketed labels such as [SSN REDACTED]
• The scrubbed (de-identified) text is what gets sent to the LLM provider — the LLM never sees raw PHI
• The original PHI string goes out of scope and is garbage-collected — no write occurs
• Our usage log records only: byte counts, token counts, processing time, redaction pattern types, and redaction count — no PHI values

Business Associate Agreement (BAA)

RGX Systems will execute a Business Associate Agreement with any VAR whose clients operate in the healthcare sector. Render Services, Inc. (our infrastructure provider) has executed a HIPAA BAA with RGX Systems, establishing the required chain of agreements. To request a BAA, contact legal@rgxsystems.com or visit /baa.

PII Scrubbing — Professional Services

When config.industry: "professional_services" is set, our scrubPII() function additionally redacts EINs, SSNs, account numbers, routing numbers, and payment card numbers before any LLM call. Same zero-persistence guarantee applies.

Technical Reference

For a full technical description of data movement through each processing stage, see the RGX Data Flow Diagram.

5. Non-Training Data Policy

RGX Systems does not use VAR payload data to train machine learning models. Specifically:

• Input text submitted to /api/v1/process or /api/v1/ingest is not stored in a form that could be used for model training
• We do not share payload data with any third party for training purposes
• Our LLM providers (Anthropic, OpenAI) operate under data processing agreements that explicitly prohibit use of API-submitted data for model training
• Usage metadata (token counts, request counts, etc.) that we do retain contains no message content and cannot be used to reconstruct submitted inputs

6. Pipeline Processing

When you call the /process endpoint with gateway routing enabled (by providing config.industry and config.routing_profile), input data is forwarded to our LLM provider under a data processing agreement that prohibits use of submitted data for model training. For healthcare and professional services, PHI/PII is scrubbed before this forwarding occurs. If you use passthrough: true, data is logged for billing purposes and not forwarded to any LLM.

7. Data Storage and Security

• All data is stored on secured cloud infrastructure (Render, PostgreSQL)
• API keys are SHA-256 hashed before storage — we cannot retrieve your plaintext key
• Sensitive configuration data is encrypted at rest using AES-256-GCM
• All data is transmitted over TLS 1.3
• Access to production systems is restricted to authorised RGX Systems personnel only
• Usage event logs are retained for billing and audit purposes for 12 months, then purged

9. Subprocessors and Data Sharing

We share data only with the following trusted subprocessors necessary to operate our infrastructure. Each subprocessor operates under a data processing agreement with RGX Systems:

• Anthropic, PBC — LLM provider. Receives only scrubbed/de-identified text (no raw PHI/PII). Prohibited from using submitted data for model training. HIPAA-compliant data processing agreement in place.
• OpenAI, LLC — Secondary LLM provider (ultra-low-latency routing profile). Same restrictions as Anthropic — no raw PHI, no model training on submitted data.
• Render Services, Inc. — Cloud infrastructure and PostgreSQL database hosting. SOC 2 Type II certified. HIPAA BAA executed with RGX Systems.
• Stripe, Inc. — Payment processing for monthly invoicing. We never store card details; Stripe is PCI-DSS Level 1 certified.

We do not share your data with any other third parties, advertisers, or data brokers. Our full subprocessor list is available on request at security@rgxsystems.com.

10. Data Retention

• PHI: Zero retention. Raw PHI is never written to storage (scrubbed in-memory before any persistence).
• VAR account data is retained for the duration of your active node access
• API usage event logs (no PHI) are retained for 12 months for billing and audit purposes, then purged
• Monthly billing aggregates are retained for 7 years as required by applicable financial regulations
• Upon termination or node revocation, account data is deleted within 30 days upon written request

11. Seat-Metering Audit Rights

What Is Metered

The Seat Metering Ledger (GET /api/v1/usage, GET /api/v1/billing/summary) records the following per tenant workspace:

• seat_id — the unique identifier for each active user seat within a tenant
• first_seen — UTC timestamp of first authenticated request for the seat this billing period
• last_seen — UTC timestamp of most recent authenticated request
• request_count — total API requests from this seat during the billing period
• active — boolean; true if the seat had at least one authenticated request this calendar month

VAR Audit API Access

VARs may query their seat ledger at any time:

• GET /api/v1/usage?month=YYYY-MM — full seat-level breakdown for any historical month (up to 12 months)
• GET /api/v1/billing/summary — current month totals: active_seats, base_fee, seat_fee, estimated_total
• GET /api/v1/clients/:ref/usage — per-tenant seat activity for a specific workspace

All billing responses are returned as structured JSON and may be downloaded by VARs for independent verification. Responses include a billing_period, seat_count, per_seat_rate, and a line_items array with one entry per active seat.

Dispute Resolution

If you believe your seat count is incorrect, contact billing@rgxsystems.com within 14 days of invoice with the disputed month and your independently counted seat list. RGX Systems will respond within 5 business days with a line-item billing export. If the export confirms a metering error, a credit will be applied to the following invoice. RGX Systems will not charge a disputed seat without providing log-level evidence of authenticated activity.

12. Your Rights

As a VAR operator, you have the right to:

• Access the data we hold about your VAR node and usage history via the /usage endpoint or by contacting us
• Request correction of inaccurate account information
• Request deletion of your account data (subject to billing retention requirements)
• Request a copy of your data in a portable format
• Withdraw consent and terminate your VAR access at any time
• Exercise seat-metering audit rights as described in Section 11

To exercise any of these rights, contact us at privacy@rgxsystems.com.

13. Cookies

Our public website uses minimal cookies necessary for session management. The VAR API does not use cookies — all authentication is via the X-Api-Key header. We do not use tracking cookies or third-party advertising cookies.

14. Business-to-Business Service

RGX Systems is a strictly business-to-business infrastructure service. We do not knowingly collect or process data from individuals under 18, and we do not provide consumer-facing products or services.

15. Changes to This Policy

We may update this policy from time to time. We will notify VAR operators of material changes via email. Continued use of the infrastructure after changes constitutes acceptance of the updated policy.

16. Contact

For privacy-related questions or requests, contact us at:

Email: privacy@rgxsystems.com
Website: rgxsystems.com