Verity is a forensic verification AI that analyzes contradictory evidence across multiple formats (video, audio, documents, images) to reconstruct verified timelines and detect inconsistencies. Built with Gemini 3's advanced reasoning capabilities and transparent thinking mode.

Fine-tuned roberta-base classifier on the LIAR dataset. Aaccepts multiple input types text, URLs, and PDFs and outputs a prediction with a confidence score. It also leverages google/flan-t5-base to generate explanations and uses an Agentic AI with LangGraph to orchestrate agents for planning, retrieval, execution, fallback, and reasoning.

ΑΩ+: Speculative reasoning verification for LLMs. Models reasoning as a potential field over token embeddings, directly linking to attention geometry. Implements tetralectic logic & Φ-harmonic stability. Efficient batched implementation with Hutchinson trace estimation, masking for variable-length sequences, and MPS/CPU hybrid for Apple Silicon.

VeritySwarm — AI truth-verification swarm. 6 agents, 2-round cross-pollination, Grok synthesis. verityswarm.com

AI research for bias mitigation and truth verification - building fair, transparent systems that combat misinformation while ensuring equitable outcomes across all demographics.

An alpha implementation of of a document format and "truth engine" designed to promote trustworthy content

Multi-model AI tribunal — consensus-driven truth verification. FastAPI + SwiftUI + Next.js + Rust.

TruthGit — Governance layer for AI agents. Consensus tracking with ontological classification.

A desktop tool for creating Bitcoin-anchored digital proofs of existence. It grounds truth in thermodynamic necessity — making deception irreversibly costly and epistemic integrity the minimum-energy state.