Podium AI is a Flask-based web application that leverages Computer Vision and multiple Generative AI models (Google Gemini, OpenAI, and Groq) to analyze and classify uploaded media.
This application serves as an AI-powered engine that processes user-uploaded images/videos using a custom Computer Vision engine (cv_engine.py) and passes the data through an AI classifier (ai_classifier.py).
Key Capabilities:
- Multi-Model Support: Integrates with Google Generative AI, OpenAI, and Groq for flexible LLM processing.
- Computer Vision Pipeline: Uses
opencv-pythonandscikit-imagefor pre-processing and feature extraction. - Web Interface: A lightweight Flask frontend allowing users to upload files for analysis.
- Backend: Python, Flask
- Computer Vision: OpenCV (
opencv-python-headless), Scikit-Image, Imutils, Numpy - GenAI Providers: *
google-generativeai(Gemini)openai(GPT models)groq(High-speed inference)
- Utilities: Python-Dotenv (Environment management)
PODIUM_AI/
├── static/ # CSS/JS files
├── templates/ # HTML Templates (Jinja2)
├── uploads/ # Directory for user-uploaded media
├── ai_classifier.py # Logic for LLM interaction/classification
├── app.py # Main Flask application entry point
├── cv_engine.py # Core Computer Vision processing logic
├── check_models.py # Utility script to verify model availability
├── requirements.txt # Project dependencies
└── .env # Environment variables (API Keys)