Skip to main content
This page provides an overview of Veratrace’s core concepts and mental model.

​
Multi-Tenant Architecture

Veratrace operates as a multi-tenant system. Each Instance is an isolated container with:
  • Its own AWS Cognito user pool
  • Separate users, TWU models, agents, and integrations
  • Independent configuration
Source: Feature Inventory - Core Objects (Instance)

​
Core Objects

​
TWU Model

A TWU Model defines the structure for Trusted Work Units. It specifies:
  • Entities - Objects that work operates on
  • Actions - Operations performed
  • Events - State changes or occurrences
  • Outcomes - Results with expressions
Models have versions with statuses: DRAFT, PUBLISHED, ARCHIVED. Source: Feature Inventory - Core Objects (TWU Model)

​
Agent

An Agent represents either an AI or human agent that performs work operations. Agents have:
  • Type (AI or HUMAN)
  • Capabilities (varies by type)
  • AI model configuration (for AI agents)
  • Priority level
Source: Feature Inventory - Core Objects (Agent)

​
Integration

An Integration connects Veratrace to external systems for evidence ingestion and data synchronization. Supports 40+ integration types across multiple categories. Source: Feature Inventory - Core Objects (Integration)

​
Work Ledger Entry

A Work Ledger Entry records a TWU execution/work operation. It tracks:
  • Execution status
  • Policy compliance status
  • Actor types (AI/HUMAN)
  • Cost and time metrics
  • Related entities
Source: Feature Inventory - Core Objects (Work Ledger Entry)

​
Mental Model

  1. Define Structure - Create TWU models that define entities, actions, events, and outcomes
  2. Configure Agents - Set up AI or human agents that will perform work
  3. Connect Systems - Integrate external systems for evidence ingestion
  4. Track Work - View work ledger entries to see actual executions, compliance, and metrics