Simple demo of how Vector Embeddings and Retrieval work in AI powered applications.
🚀 You can check out a working demo here: DEMO
Vector Embeddings are crucial to the way AI Systems remember things. You'll see in this demo how the computer is able to convert plain text into semantic groupings and find sections by ideas!
The powerful building blocks of AI retrieval systems are simple, transparent and open to anyone. You should be aware of how these systems work and not be a blind user!
Clone this repo and find out how it works under the hood.
git clone https://github.com/bculleec/Vector-Embeddings-Retrieval-in-Node.js
cd Vector-Embeddings-Retrieval-in-Node.jsnpm installDrop any .txt file inside the data/ directory.
The included example uses Alice in Wonderland, but you can replace it with anything you like.
node main.jsThis runs your pipeline to chunk the text and create:
chunks.json→ all the text chunksvectors.json→ vector representations of those chunks
npm startor
node server.jsThis serves both the backend API and the frontend demo at http://localhost:3000.