New Text Document.txt

🛠️ RoomMakeover.AI – Tech Stack Documentation

Project Goal:
Create an AI-powered platform where users upload a picture of a room and receive decor improvement suggestions (with product links) based on a given style and budget — using Computer Vision and LLMs.
🔧 Finalized Tech Stack
🧠 1. Large Language Model (LLM) Reasoning
Component	Technology Used	Purpose
LLM Backend	Gemini 1.5 Flash	Fast, cost-effective reasoning engine for decor suggestions
Prompt Chaining	LangChain	Manage multi-step reasoning and prompt flows
LLM Tasks	Suggest decor items, budget planning, product matching, style analysis	
🧍‍♂️ 2. Computer Vision (Room Image Analysis)
Task	Technology/Library Used	Description
Object Detection	YOLOv8 (via Ultralytics)	Identify existing room elements (e.g., bed, table, windows)
Region Segmentation (optional)	Segment Anything Model (SAM)	Highlight areas where new items can be placed
Aesthetic/Style Feature Extraction	CLIP	Match room image with decor styles (boho, minimalist, etc.)
🔍 3. Product Recommendation Engine
Component	Technology / Approach	Purpose
Product Search (optional)	SerpAPI (Google/Amazon), Amazon Product Advertising API	Get real-time product info & links
Mock Product Dataset	Custom JSON with images, prices, links	Useful for MVP/demo without external APIs
Budget Filtering Logic	Custom LangChain prompt or Python logic	Ensure suggestions stay under budget
🖥️ 4. Frontend Interface
Platform	Technology Used	Purpose
Web Frontend	Streamlit	Quick, interactive UI to upload image and receive suggestions
Alternate UI	FastAPI + React (optional)	Modular backend + modern frontend (future-proofing)
User Input	Image upload, budget entry, style selector	
⚙️ 5. Backend & Deployment
Component	Technology Used	Purpose
API Backend (if used)	FastAPI	For LLM pipeline, CV inference, image handling
Storage	Local or Cloudinary	To host uploaded room images
Hosting (for demo)	Hugging Face Spaces, Render, Replit, Vercel	Free-to-use cloud hosting for GenAI projects
API Key Management	.env + os.environ	Keep LLM & CV API keys safe and secure
🔒 6. Security & Privacy Considerations
Area	Security Practice
Image Uploads	Input validation, max size control
API Keys	Stored securely in .env file
User Data	No persistent data storage unless user opts in
API Usage	Rate limits, fallback modes to avoid abuse