Skip to content

Getting Started with Lucid

Welcome to the Lucid Developer Platform! Lucid helps you build Secure AI Supply Chains by enforcing safety policies (Auditors) within Trusted Execution Environments (TEEs).

This guide will take you from installation to deploying your first hardware-secured AI workload.

🚀 The Path to Production

To get started with Lucid, we recommend following these steps in order:

  1. Installation: Set up the Lucid CLI and SDK on your local machine.
  2. Cluster Setup: Initialize a local development environment (Kind) or connect to a cloud TEE cluster.
  3. Your First Auditor: Walk through building, verifying, and deploying a simple PII-redaction auditor.

🏗️ Core Concepts

Before diving in, it may be helpful to understand the high-level architecture:

  • Auditor Chain: A sequence of safety sidecars that inspect model input and output.
  • Confidential Computing: The hardware-based security layer that protects your models.
  • AI Passport: The cryptographic proof of compliance generated for every request.

[!TIP] Need a faster look? Check out the Interactive Demo in the repository to run the entire system with one command.