VerifAI

A small toolkit that detects likely AI-generated content from videos, images and text.

How to run

These instructions assume you're on Windows (developer environment) and have Python 3.10+ and Node.js installed.

• Create a Python virtual environment and install backend dependencies:

python -m venv .venv
.\.venv\Scripts\Activate.ps1   # PowerShell
pip install -r backend/requirements.txt

• Start the FastAPI backend (serves the analysis API):

# from repository root
cd backend
uvicorn app.main:app --reload --port 8000

• Start the auxiliary Flask server (uses Google GenAI integration):

cd server
# ensure GEMINI_API_KEY is set in your environment
$env:GEMINI_API_KEY = "<your_key>"
python app.py
# Flask defaults to port 5000 (see server/app.py)

• Frontend (web) — run the Vite app(s):

cd web
npm install
npm run dev
# default Vite port is 5173

cd ../extension
npm install
npm run dev

• Running tests (Python):

# from repo root
pip install pytest
pytest -q

Backend

cd backend/app

🏗️ Architecture

A simplified visual of how requests flow through the project (Mermaid):

flowchart LR
	subgraph Frontends
		Web["Web UI (Vite)"]
		Ext["Browser Extension"]
	end

	Frontends -->|POST /analyze| BackendAPI["Backend (FastAPI)\nuvicorn :8000"]

	subgraph Processing
		BackendAPI --> Core["Core logic (app.core)"]
		Core -->|video -> yt-dlp / ffmpeg| VideoWorker["Video fetch & frames\n(yt-dlp, OpenCV)"]
		Core -->|image -> OCR/vision| ImageWorker["Image analysis (Pillow, OpenCV)"]
		Core -->|text -> LLM| LLMProxy["Google GenAI / local model"]
	end

	Core -->|calls| Server["Auxiliary server (Flask)\nGoogle GenAI wrapper :5000"]
	LLMProxy --- Server

	VideoWorker --> Core
	ImageWorker --> Core

	BackendAPI -->|response| Frontends

	classDef infra fill:#f9f,stroke:#333,stroke-width:1px;
	class BackendAPI,Server,Core infra;

Loading

Notes:

• The FastAPI service exposes /analyze which accepts either url (video) or text (text analysis).
• The server app (Flask) wraps calls to the Google GenAI SDK; it expects GEMINI_API_KEY in env.

Important files

• Backend: backend/app/main.py — FastAPI entrypoint
• Server: server/app.py — Flask/GenAI helper
• Frontend (web): web/
• Extension UI: extension/
• Tests: tests/

Environment variables

• GEMINI_API_KEY — required for the Google GenAI integration used by server/app.py.

Developer notes & tips

• The backend chooses analysis type by the incoming JSON: provide url to analyze videos or text to analyze text (see backend/app/main.py).
• Video processing uses yt-dlp (see backend/requirements.txt) — ensure ffmpeg is installed on your PATH for frame extraction.
• Redis and fakeredis are listed in requirements.txt for caching/testing; configure a real Redis instance via environment variables if needed.

---