BreedSense AI (SIH 2025 By Gradient Gang)

BreedSense AI — in-browser cattle/buffalo breed recognition (Vite + React + TypeScript)

Overview

Frontend captures or uploads images, sends them to a FastAPI backend for breed recognition, and displays a normalized breed profile (Overview, Gender Traits, Production (peak milk), History, Conservation). Identifications are persisted to localStorage.

Tech stack

• Frontend: Vite, React, TypeScript, Tailwind CSS, Radix UI
• Backend: FastAPI (expected endpoints below)
• Local persistence: localStorage (key: vxai_identifications_v1)

Required backend endpoints

• GET / → root ping
• POST /recognize_breed → multipart/form-data (file)
  • Response: { predicted_class: string, confidence_score: number } (200) or 400 with { detail: "No Buffalo Detected..." }
• GET /buffalo_breeds/ → returns list of breed names
• GET /buffalo_breeds/{breed_name} → returns detailed breed JSON (the frontend normalizer tolerates many shapes, including nested male/female fields)

Environment

Create a .env in the project root (my-app) or set env before running:

• VITE_API_BASE — base URL for your FastAPI server (default fallback: http://localhost:8000)