Documentation

Documentation

How Salus detects, tokenises and restores sensitive data inside your environment, and how to integrate, deploy and operate it. Written against the current release.

What is Salus

Getting started

What Salus is, how a request travels through it, and how a first deployment runs.

  1. What is Salus
  2. How Salus works
  3. Where Salus sits
  4. First deployment

Integrate

Call Salus from your applications, your AI gateway or your batch jobs. The API, contexts, policy and restore.

  1. Overview
  2. Detect text
  3. Documents and images
  4. Contexts and restore
  5. LiteLLM guardrail
  6. Keys and tenants

Endpoints

Salus on managed devices for the AI applications you cannot change. Files, images and review before send.

  1. Overview

Core concepts

Tokenisation, masking, pseudonymisation and why relationships survive.

  1. Tokenisation
  2. Versus masking
  3. Versus pseudonymisation
  4. Anonymisation
  5. Context preservation

Architecture

The services, the detection stack, the context store, the restore path, human review and vision.

  1. Services
  2. Detection pipeline
  3. Detection classes
  4. Vault
  5. Restore
  6. Human review
  7. OCR and images

Deployment and operations

OpenShift and Compose installs, sizing, configuration, the console, audit, metrics and releases.

  1. Overview
  2. OpenShift
  3. Docker Compose
  4. Configuration
  5. Console
  6. Audit and logging
  7. Releases

FAQ

Direct answers to the questions security and architecture teams ask first.

  1. Technical FAQ

All example values in this documentation are synthetic.