Core Concepts

Meridian is easiest to use once the underlying mental model is clear. This section explains how the product thinks about evidence, configuration sources, runtime patching, engines, and artifacts.

Read this section if you need to understand

• what a Meridian result actually means
• why some config inputs can be parsed locally and others cannot
• why safe mode exists and how it differs from tee and live
• how engine resolution affects both semantic validation and runtime execution
• why artifact provenance matters in review

Recommended order

1. Evidence Model
2. Config Source Model
3. Semantic Validation
4. Runtime Modes and Patching
5. Runtime Engine Model
6. Artifact Model