Research Project

Detection and Risk Assessment of Hallucinations in Generative AI Systems

A browser-assisted verification workflow for AI-generated content. The system combines a Chrome extension with a FastAPI backend to inspect responses, score trust, flag suspicious claims, and keep lightweight analysis history.

View Repository Read Documentation Download ZIP

System Components

  • Chrome extension for supported AI chat platforms
  • FastAPI backend for scoring and summarizing hallucination risk
  • SQLite-backed recent analysis history

Supported Platforms

  • ChatGPT
  • Gemini
  • Claude
  • Perplexity

API Endpoints

  • /health for uptime checks
  • /analyze for response scoring
  • /history for recent saved analyses

Download Source

Grab the complete repository as a ZIP archive for local setup, demo packaging, or review.

Full Project ZIP

Browse Extension

Open the browser-extension code directly if you only want the Chrome-side implementation.

Open Extension Folder

Browse Backend

Inspect the FastAPI service, analyzer logic, verification layer, and storage code.

Open Backend Folder

Download the code directly from this page, then host the backend on Render and use the extension locally or publish it separately.