// DATA GOVERNANCE & SOVEREIGNTY

Privacy Policy & Data Sovereignty

EFFECTIVE DATE: NOVEMBER 4, 2024 • VERSION 2.0.0

1. Architectural Foundations: Local-First Data Sovereignty

Relay is engineered as an organizational continuity platform built to run within your organization’s controlled compute infrastructure. Unlike conventional SaaS products that ingest confidential corporate archives into centralized multi-tenant databases, Relay deploys via containerized environments (Docker) and persists all knowledge locally using an isolated SQLite database engine with Write-Ahead Logging (WAL) and strict foreign-key enforcement.

Your institutional documents, board minutes, personnel records, regulatory audits, and responsibility graphs reside on your designated persistent storage mounts. Relay Systems does not host, mirror, or inspect your primary organizational records.

2. Information Processing & Document Extraction

When you stage records in the Relay Knowledge Archive (such as PDF, DOCX, or TXT files), processing occurs locally within the backend container.

  • Client-Side Validation: Files are validated for size (up to 25MB) and MIME type directly in your browser session before submission.
  • Cryptographic Integrity: Ingested files receive SHA-256 hashes to guarantee data provenance and prevent tampering.
  • Entity & Citation Triples: Verbatim text excerpts are indexed strictly for organizational citation and handover cross-referencing.

3. Artificial Intelligence & Model Invocation Boundaries

When Relay’s specialized continuity agents (Knowledge Extractor, Gap Detective, Continuity Analyst, Handover Architect) execute analytical pipelines:

  • Organizational text passages are transmitted solely for ephemeral contextual inference.
  • Zero customer data is ever permitted to be used for commercial model training, tuning, or evaluation by external LLM vendors.
  • Self-hosted or air-gapped LLM inference backends can be configured in your environment variables without modifying application business logic.

4. Telemetry, Tracking, and Cookies

Relay does not implement third-party tracking scripts, marketing pixels, behavioral telemetry, or commercial advertising analytics. We believe organizational transitions involve the most sensitive corporate assets; tracking user behavior inside an institutional archive is incompatible with our security principles.

5. Data Retention, Purging, and Deletion

Because Relay stores data on your local host volume (e.g. backend/data/relay.db), your administrators retain complete, unilateral authority over data retention. You may prune documents, archive transition cases, or purge the database volume entirely using standard container commands:

docker compose down -v