Documentation
Orthalon

Orthalon Documentation

Set up Orthalon, protect workforce AI use, and operate browser, desktop, API, and MCP coverage from one centralized console.

Getting Started: Orthalon stores metadata by default.
Enterprise teams can optionally retain encrypted post-redaction evidence without the original detected values.


1. Quick Navigation


2. Protected Environments Matrix

EnvironmentPrimary TargetProtection ScopeStarting Plan
Browser Extension40+ Web AI Portals (ChatGPT, Claude, Gemini)Prompt DLP, coaching modals, file upload inspectionMonitor
Desktop ProtectionLocal IDEs (VS Code, Cursor) & CLI toolsLocal model API traffic, CLI commandsMonitor
AI GatewayServer-side microservices & LLM backendsReverse proxy for OpenAI, Anthropic, Gemini API trafficMonitor
MCP Security AgentClaude Desktop, Cursor MCP servers, stdio toolsMCP server inventory, tool-call parameter redactionEnterprise
Discovery ConnectorsIdentity providers, CASB, Secure Web GatewaysAsset inventory and shadow AI usage correlationEnterprise

3. Privacy & Zero-Retention Model

  • Local Endpoint Inspection: For browser extensions and desktop agents, detection and redaction execute locally in memory on the employee workstation.
  • Metadata by Default: Orthalon normally ingests sanitized security telemetry rather than prompt text. Enterprise administrators can explicitly retain encrypted post-redaction text with typed placeholders and bounded retention.
  • Deterministic Policy Bundles: Sensors poll compiled immutable policy bundles that evaluate with sub-millisecond local latency.