Getting Started with TraceMem
TraceMem is the system of record for AI decisions. It captures every decision made by AI agents—including what data was used, which policies were checked, who approved exceptions, and what changed as a result—creating an immutable audit trail that answers "Why was this allowed?"
Key Insight: Logs show what happened. TraceMem shows why it happened.
What is TraceMem?
TraceMem provides explainable and auditable AI agent workflows by:
- Capturing decision context: Every decision is wrapped in a Decision Envelope that tracks intent, automation mode, and lifecycle state
- Governing data access: Agents access data through Data Products that define what's exposed, for what purposes, and under which restrictions
- Enforcing policies: Deterministic rules determine whether actions are allowed, denied, or require human approval
- Recording approvals: Human judgment is captured as durable evidence when policies require exceptions
- Creating immutable traces: Every decision creates a complete, queryable audit trail
All of this happens automatically as your agents interact with TraceMem through the Agent MCP (Model Context Protocol) server.
Quick Start
- Quickstart Guide - Get up and running in 5 minutes
- Installation - Install and configure TraceMem
- Setup Guide - Complete setup walkthrough
Core Concepts
To understand TraceMem, familiarize yourself with these core concepts:
- Decision Traces - Immutable records of what happened and why
- Approvals - Human judgment as durable decision evidence
- Agents - AI systems that make decisions through TraceMem
- Data Products - Governed, purpose-bound interfaces to your data
- Policies - Deterministic decision rules
- Approval Routes - Approval delivery configuration
- Connectors - Connections to external data systems
- Integrations - Approval delivery integrations