Crate truthlens 

TruthLens — AI Hallucination Detector

Analyzes AI-generated text for hallucination risk using linguistic signals, claim decomposition, and formally verified trust scoring.

Quick Start

use truthlens::{analyze, TrustReport};

let text = "Albert Einstein was born in 1879 in Ulm, Germany.";
let report = analyze(text);
println!("Trust: {:.0}% — {}", report.score * 100.0, report.risk_level);

Re-exports

pub use claim::extract_claims;

pub use claim::Claim;

pub use consistency::check_consistency;

pub use consistency::ConsistencyReport;

pub use consistency::Contradiction;

pub use scorer::score_claim;

pub use scorer::score_passage;

pub use scorer::RiskLevel;

pub use scorer::TrustScore;

pub use trajectory::analyze_trajectory;

pub use trajectory::TrajectoryAnalysis;

pub use trajectory::TrajectoryPattern;

Modules

claim

consistency

entity

Entity Cross-Reference with Wikidata SPARQL (v0.4)

scorer

trajectory

Structs

ClaimReport

Analysis report for a single claim.

TrustReport

Full analysis report for a text passage.

Functions

analyze

Analyze a text passage for hallucination risk.

analyze_with_verification

Analyze text with entity verification against Wikidata.