Installation
Get the Lucid CLI installed and ready to use.
Alpha Access Required
The Lucid CLI and SDK are available to alpha participants. Request access to get started.
Quick Install
That's it! Verify everything works:
Core Commands:
apply Deploy environment (serverless or self-hosted)
diff Preview changes before applying
status View agent status
logs Stream agent logs
start Start a stopped agent
stop Stop a running agent
teardown Clean up environment
export Export cluster state to YAML
Serverless Mode:
lucid apply --app open-webui --model llama-3.1-8b
lucid catalog models List available models
lucid catalog auditors List auditor profiles
lucid verify endpoint Verify TEE attestation
Other Commands:
login Authenticate with Lucid platform
catalog Browse resource catalog (apps, models, auditors)
passport View AI passports
auditor Manage Lucid Auditors
verify Verify TEE attestation
Authenticate
Connect to the Lucid platform:
Password:
Logged in as user@example.com
For automation and CI/CD, use the LUCID_API_KEY environment variable:
# Using API key (recommended for CI/CD)
export LUCID_API_KEY=your-api-key
lucid status # commands will use the API key automatically
Additional login options:
# Login with email (will prompt for password)
lucid login -e user@example.com
# Generate an API key after login
lucid login --generate-key
Credentials are stored in ~/.lucid/config.yaml.
Troubleshooting
Python Version Issues
Error: Python version 3.x is not supported
Lucid CLI requires Python 3.12 or higher.
If you have multiple Python versions, use a specific one:
python3.12 -m pip install lucid-cli
Or use a tool like pyenv to manage versions.
pip SSL Errors
Error: SSL: CERTIFICATE_VERIFY_FAILED or similar SSL errors
-
Update pip and certifi:
pip install --upgrade pip certifi -
On macOS, run the certificate installer:
/Applications/Python\ 3.12/Install\ Certificates.command -
Behind a corporate proxy, you may need to configure certificates:
pip install --trusted-host pypi.org --trusted-host github.com lucid-cli
Prerequisites
Before using lucid apply, ensure you have:
- Python 3.12+
- Docker Desktop (or equivalent OCI runtime) - for local development
Next Steps
You're ready to go!
- Set up your cluster - Configure where your agents run
- Deploy an agent - Get an AI workload running with auditing
Building Custom Auditors?
If you're developing custom safety auditors, you'll also need the Lucid SDK:
pip install lucid-sdk
See the Auditor Development Guide for details.