Lucid Developer Platform

Lucid is a developer-first platform for building Secure AI Supply Chains. It transforms traditional opaque AI systems into "Glass Box" environments, ensuring data privacy and model integrity through cryptographic hardware-based attestation.

⚡ Quick Start

1. Install Lucid CLI

Install the Lucid CLI to manage your clusters, auditors, and deployments.

pip install -e packages/lucid-cli

2. Define & Deploy

Define your Audit Chain as policy-as-code. The Lucid CLI then automatically injects these safety sidecars into your standard Kubernetes manifests at deploy-time.

# auditors.yaml
chain:
  - name: pii-scanner
    description: "Scan for SSN and other PII in prompts"
    image: lucid-auditor-sidecar:latest
    script: pii_auditor.py
    port: 8081

  - name: injection-detector
    description: "Detect prompt injection attacks"
    image: lucid-auditor-sidecar:latest
    script: injection_auditor.py
    port: 8082

Deploy your manifest:

lucid deploy apply --file manifest.yaml

3. Observe

Monitor live compliance and audit logs via the Lucid Observer. Every decision made by your auditors is recorded with cryptographic proof.

Follow the Audit logs using the Lucid CLI:

lucid logs view --follow

Or through the observer dashboard:

4. Verify & Integrate (AI Passport)

Every inference generates a cryptographically signed AI Passport. Integrate this certificate into your downstream applications to prove compliance to the users of your AI applications.

import requests

# Retrieve the AI Passport for a specific session
response = requests.get("http://verifier-service:8000/passport/session-123")
passport = response.json()

if passport["deployment_authorized"]:
    print("✓ Cryptographically Verified: Policy Enforced in TEE")

This is what the AI Passport will look like to the user of your AI application:

� Getting Started

• Installation: Set up the Lucid CLI and SDK.
• First Auditor: Build and deploy your first safety node in 5 minutes.
• Cluster Setup: Make your Kubernetes cluster Lucid-ready.

�📖 Key Concepts

• Glass Box Philosophy: Turning opaque infrastructure into verifiable execution.
• Phase-Aware Audit: Chaining guardrails across Build, Input, Runtime, and Output.
• Confidential Computing: Leveraging hardware-based isolation (Intel SGX, AMD SEV).

� Developer Guides

• Auditor Development: Master the SDK patterns and lifecycle hooks.
• Policy as Code: Define your safety guardrails with auditors.yaml.
• Deployment Workflow: Zero-touch TEE injection for any K8s workload.

🛠 Reference & API

• Lucid CLI Reference: Command-line interface documentation.
• Lucid SDK Reference: Python library API documentation.
• SaaS Verifier API: Documentation for the hosted Lucid Verifier API.