TruthLens
AI Hallucination Detector — Formally verified trust scoring for LLM outputs. Analyze AI-generated text for hallucination risk with no API keys, no LLM calls. Fast, local, and formally verified in Lean 4.
How to pay
Subscribe
$5/month
Predictable monthly cost with included usage. Best for steady, high-volume traffic.
- Unlimited tools within plan limits
- One API key, billed once a month
- Cancel any time
TruthLens — AI Hallucination Detector
Formally verified trust scoring for LLM outputs. Analyze AI-generated text for hallucination risk with no API keys, no LLM calls. Fast, local, and formally verified in Lean 4.
How It Works
TruthLens decomposes AI-generated text into atomic, independently verifiable claims and evaluates them across multiple signals — confidence, hedging (uncertainty markers), specificity, verifiability, and consistency. These signals are aggregated into a trust score on a [0.0, 1.0] scale, ranging from LOW (factual/hedged) to CRITICAL (hallucination risk).
Key Features
- No LLM/API Required: Works entirely locally and offline
- Formal Verification: Lean 4 proofs guarantee score bounds, monotonicity, and functional purity
- Consistency Checking: Detects contradictions between multiple AI responses
- Entity Verification: Optional cross-referencing against Wikidata
- Flexible Integration: CLI, Rust/Python library, and MCP server
Installation
Via MCPize: mcpize install truthlens
As CLI: cargo install truthlens
Python library: pip install truthlens
Snap: sudo snap install truthlens