🤖** AI Chatbot with Pinecone + OpenAI (RAG System)**

A smart AI chatbot built using Node.js, OpenAI embeddings, and Pinecone vector database. It uses a Retrieval-Augmented Generation (RAG) approach to answer questions based on custom data.

🌿** Features** 📄 custom .txt data 🧠 Convert text into embeddings using OpenAI 📦 Store embeddings in Pinecone vector DB 🔍 Semantic search using user queries 🤖 AI-generated responses using OpenAI GPT models 🧩 Modular backend architecture (controllers, services, utils)

🏗️** Tech Stack** Node.js Express.js OpenAI API (Embeddings + Chat Completions) Pinecone Vector Database JavaScript (ES Modules)

📁 Project Structure

src/ │

├── config/ │ ├── openai.js │ └── pinconeapi.js │

├── controllers/ │ ├── chatController.js │ └── adminController.js │

├── services/ │ ├── embeddingServices.js │ ├── openaiServices.js │ └── pineconeService.js │

├── utils/ │ └── searchContext.js │

├── routes/ │ ├── adminRoutes.js │ └── chatRoutes.js │

├── data/ │ ├── about.txt │ └── skills.txt │

└── scripts/ (optional used for testing purpose only) └── uploadData.js

⚙️ How It Works (RAG Flow) User Question

↓

Generate Embedding (OpenAI)

↓

Search Similar Vectors (Pinecone)

↓

Retrieve Relevant Context

↓

Send Context + Question to GPT

↓

AI Response

⚡ Performance Notes Embeddings are generated using OpenAI API

Pinecone handles fast semantic search

Response time: ~2–4 seconds (normal RAG behavior)

🧠 Key Concepts Used Embeddings (vector representation of text)

Semantic Search

Vector Database (Pinecone)

RAG Architecture

🤍 Author

Kinza AI-Focused MERN Stack Developer