AletheionGuard

Documentation

Learn how to integrate AletheionGuard into your AI applications for epistemic uncertainty quantification.

Quick Start

Get started with AletheionGuard in 5 minutes

# Install the package
pip install aletheion-guard
# Quick example
from
aletheion_guard
import
EpistemicAuditor
auditor = EpistemicAuditor()
result = auditor.evaluate(
"The capital is Paris"
)
print
(result.verdict)
# "ACCEPT" | "MAYBE" | "REFUSED"

Key Features

🎯 Q1 & Q2 Separation

Distinguish between aleatoric (data noise) and epistemic (model ignorance) uncertainty

📐 Pyramid Height

Unified metric combining Q1 and Q2 to measure proximity to truth

🏆 TIER GOLD

Meets highest standards for epistemic uncertainty quantification

⚡ Fast API

Production-ready FastAPI with <50ms latency

🔧 Model Agnostic

Works with any LLM: GPT-4, Claude, Llama, Mistral, and more

📊 Calibration Metrics

ECE, Brier score, and custom calibration metrics included

Use Cases

Enterprise LLM Safety

Audit GPT-4 responses before showing to customers. Prevent hallucinations in production systems.

Healthcare & Legal AI

High-stakes domains requiring uncertainty quantification. Meet regulatory requirements for AI transparency.

RAG Optimization

Detect when LLM needs more context. Trigger additional retrieval when epistemic uncertainty (Q2) is high.

Model Research

Compare calibration across models. Identify which models are better calibrated for your domain.

Ready to Get Started?

Create a free account and get 1,000 requests per month

Need help? Check out our Quick Start Guide or contact support