Agenrta Team: CertiFy (FactChecker and Information Verifier)
In today’s generative era—where deepfakes and AI-driven misinformation spread in minutes—FactChecker is essential infrastructure.
Core trick
Extracts claims, retrieves evidence, then has analyst and skeptic agents debate them before a judge agent issues a verdict.
Description
In today’s generative era—where deepfakes and AI-driven misinformation spread in minutes—FactChecker is essential infrastructure. Powered by GPT-5’s multi-modal reasoning, it ingests text, images, video, audio, or links; extracts factual claims; retrieves live, high-credibility evidence via hybrid search; and runs an adversarial debate between AI agents. The Analyst Agent builds the strongest pro case, the Skeptic finds contradictions, the Debating Agent moderates, and the Judge issues a True/Fake/Unverified verdict with calibrated confidence. GPT-5 enables deep reasoning, long-context handling, and clear natural-language explanations. Outputs include clickable citations, debate highlights, and visual reasoning maps, integrating seamlessly into SaaS, APIs, and moderation pipelines. As AI content explodes, every platform, enterprise, and government will require real-time verification—unlocking a multi-billion-dollar market for trustworthy, explainable safeguards.
README excerpt
# Agentra Team – CertiFy **Multi-Modal FactChecker & Information Verifier** --- ## 📌 Overview In today’s **generative AI era**—where deepfakes, synthetic media, and AI-generated misinformation can spread globally in **minutes**—truth verification must match the **speed, scale, and complexity of the internet**. **CertiFy** is a **real-time, multi-modal fact-checking engine** powered by **GPT-5’s vision-language reasoning**. It ingests any combination of: - 📝 **Text** - 🖼 **Images** - 🎥 **Video frames** - 🎙 **Audio** - 🔗 **Links** …and automatically: 1. **Extracts factual claims** 2. **Retrieves high-credibility evidence** using **hybrid search** (web APIs, news APIs, domain trust scoring, deduplication) 3. **Runs an adversarial, citation-driven debate** between AI agents 4. **Issues a transparent verdict** — *True*, *Fake*, or *Unverified* — **with calibrated confidence** --- ## 🧠 How It Works | **Agent** | **Role** | |------------------|----------| | **Analyst Agent** | Builds the strongest *pro* case for each claim | | **Skeptic Agent** | Finds contradictions, missing context, and counter-evidence | | **Debating Agent**| Moderates a turn-based exchange with rebuttals, cross-examinations, and reliability scoring | | **Judge Agent** | Delivers the final verdict with confidence scores | --- ## 🔍 - **Multi-modal parsing** → Understands text, images, and video **simultaneously** - **Advanced reasoning** → Weighs conflicting