Principle:Protectai Llm guard Scanner Configuration Schema

Knowledge Sources	LLM Guard LLM Guard API Documentation
Domains	Configuration, API_Design, DevOps
Last Updated	2026-02-14 12:00 GMT

Overview

A YAML-based declarative configuration pattern that defines scanner pipelines, authentication, rate limiting, and observability settings for the LLM Guard API server.

Description

The scanner configuration schema enables declarative definition of the entire LLM Guard API server behavior through a single YAML file. It uses Pydantic models for validation and supports environment variable substitution via ${ENV_VAR:default} syntax.

The configuration covers:

• Scanner pipelines: Ordered lists of input and output scanners with per-scanner parameters.
• Application settings: Log level, fail-fast mode, timeouts, lazy loading.
• Authentication: HTTP Bearer or HTTP Basic authentication.
• Rate limiting: Request rate limits with enable/disable toggle.
• Observability: OpenTelemetry tracing and metrics exporter configuration.

Usage

Use this principle when deploying LLM Guard as a production API service. The YAML configuration file is the single source of truth for the server's behavior, loaded at startup via the CONFIG_FILE environment variable.

Theoretical Basis

# Pseudocode for configuration loading
yaml_content = load_yaml(config_file)        # With env var substitution
config = Config(**yaml_content)              # Pydantic validation
scanners = instantiate_scanners(config)      # Create scanner instances
app = create_fastapi_app(config, scanners)   # Wire up routes

Related Pages

Implemented By

• Implementation:Protectai_Llm_guard_Config