https://github.com/kanadshee-18/vector-store

A minimal project to understand how cosine similarity works in a vector database 🧠📊. It demonstrates semantic search by converting text into embeddings and comparing them using vector math.
https://github.com/kanadshee-18/vector-store

ai cosine-similarity expressjs gemini node-cache nodejs textembedding ts-node typescript uuidv4 vector vector-search

Last synced: 8 months ago
JSON representation

A minimal project to understand how cosine similarity works in a vector database 🧠📊. It demonstrates semantic search by converting text into embeddings and comparing them using vector math.

• Host: GitHub
• URL: https://github.com/kanadshee-18/vector-store
• Owner: KanadShee-18
• License: mit
• Created: 2025-05-03T17:40:29.000Z (8 months ago)
• Default Branch: main
• Last Pushed: 2025-05-04T06:24:31.000Z (8 months ago)
• Last Synced: 2025-05-04T07:25:07.729Z (8 months ago)
• Topics: ai, cosine-similarity, expressjs, gemini, node-cache, nodejs, textembedding, ts-node, typescript, uuidv4, vector, vector-search
• Language: TypeScript
• Homepage:
• Size: 185 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# 🔍 Vector Similarity Search Demo

---

## A simple visualization:

![Cosine Similarity](cosine-similarity_visualization.png)

> A minimal project to understand how **cosine similarity** works in a **vector database** 🧠📊.  

> It demonstrates **semantic search** by converting text into embeddings and comparing them using vector math.

---

## 🧰 Tech Stack

- ⚙️ Node.js + TypeScript  

- 🧾 Custom in-memory vector store  

- 📐 Cosine similarity algorithm

---

## 💡 What You’ll Learn

- 🔤 How text embeddings are used in vector search  

- 📏 How cosine similarity compares semantic meaning  

- 🎯 Filtering using `topK` and `threshold` parameters  

- ⚡ Real-world basics of how vector DBs like Pinecone, Weaviate, or FAISS work

## License

[MIT License](LICENSE)