Skip to content

README.md

Latest commit

 

History

History
36 lines (28 loc) · 1.35 KB

README.md

File metadata and controls

36 lines (28 loc) · 1.35 KB

Apollo – Truth, Verified by AI

Apollo is an Agentic AI system that detects whether an image or video is AI-generated and explains why it believes so. It combines deepfake detection models with a reasoning agent that communicates findings in simple, human terms.

Theme

Built for MumbaiHacks 2025 – Track: Misinformation
Theme: Agentic AI

Concept

In an age where AI-generated visuals blur reality, Apollo acts as the "god of truth" — analyzing media for authenticity and providing transparent, explainable verification.

Tech Stack

  • Python
  • PyTorch / TensorFlow
  • Hugging Face (AI-content detection models)
  • Streamlit or Gradio (for the interface)
  • OpenAI API or Llama 3 (for reasoning)
  • EXIF & metadata analysis tools

How It Works

  1. User uploads an image (or short video frame).
  2. Detection model predicts authenticity confidence.
  3. Agent layer analyzes artifacts, metadata, and context.
  4. Apollo explains reasoning:
    “Lighting inconsistencies suggest AI generation (Confidence: 87%).”
  5. (Future scope) Adds browser extension + API integration.

Business Model

Freemium for users, API licensing for enterprises, partnerships with media & government agencies.

Building

  • Built by Team Helios
  • Built for MumbaiHacks 2025

Status

MVP under development – to be demoed live at MumbaiHacks 2025.