📚 AI-Search

A simple semantic search engine for PDFs using embeddings + Pinecone vector database — written in Python.

Given a set of PDF documents, this project:

✔ Loads & chunks text ✔ Generates vector embeddings ✔ Stores vectors in Pinecone ✔ Queries them with natural language ✔ Returns relevant passages, document names, and page numbers

🧩 Features

Extract text from PDF files

Compute vector embeddings via your adapter (e.g., LangChain, OpenAI)

Index vectors in a Pinecone index

Run semantic search queries

Filter and dedupe results

Stream or show exact matches + context

🚀 Quick Start Requirements

Before you begin, make sure you have:

✔ Python 3.8 or higher ✔ Pinecone API key ✔ OpenAI key (or other embedding provider)