#VectorForge

A production-inspired Vector Database built from scratch in JavaScript to understand how modern vector databases like Pinecone, Weaviate, Chroma, and Milvus work internally.

Instead of relying on existing vector database services, VectorForge implements the core search algorithms, indexing structures, distance metrics, and REST APIs from scratch, including Brute Force Search, KD-Tree, HNSW (Hierarchical Navigable Small World), Cosine Similarity, Euclidean Distance, Manhattan Distance, and a Retrieval-Augmented Generation (RAG) pipeline powered by Ollama, to demonstrate how modern semantic search systems work internally.