OCR on the server first
Heavy OCR and inspection stay centralized for patching, audit, and performance reasons, with room for offline fallback later.
Architecture
Consent Secure uses a practical enterprise model. Endpoints capture, enforce, and report. Server services handle heavy inspection, orchestration, storage, and audit. Shared services coordinate policy, identity, events, and operator workflows.
Single endpoint host for DLP communications, secure token storage, heartbeat, pending commands, and local remediation execution.
Consent, privacy, DLP, DSPM, endpoint APIs, IAM, and platform ops run as separated services with shared contracts and clean module boundaries.
RabbitMQ for messaging, Redis for presence and scale-out, PostgreSQL for data, Docker for deployment, and optional Elastic or other SIEM targets.
Unified operator UI with pages for endpoints, deployments, commands, SIEM, threats, and events backed by server-side pagination and filters.
Key Design Choices
Heavy OCR and inspection stay centralized for patching, audit, and performance reasons, with room for offline fallback later.
OCR, connector, and future AI integrations are kept behind contracts so infrastructure choices do not leak into application workflows.
RabbitMQ, Redis, SignalR, and replayable SIEM export patterns support fleet operations and operational resilience.
Layered Delivery
The architecture is easier to understand when shown as clear layers: endpoint capture, centralized services, and shared operational infrastructure.
Deployment
The current stack is Dockerized, supports VM-based deployment and GitHub Actions-based rollout, and carries documented runbooks for endpoint readiness, SIEM operations, and deployment packaging. Managed cloud service choices can be layered in later without rewriting the public architecture story.