Vector Search Demo

Semantic search with 2D embedding visualization

Back

Click to load the embedding model in your browser. No API key required.

How it works

  • • Model runs entirely in your browser via WebAssembly — no server
  • • Uses all-MiniLM-L6-v2 (32MB) for 384-dimensional semantic embeddings
  • • 2D visualization projects embeddings using PCA-style dimensionality reduction
  • • Cosine similarity ranks documents by semantic relevance
  • • Click visualization dots to search by those documents
  • • No API keys or server-side processing required

Why this matters

Semantic retrieval is foundational for enterprise support, search, and knowledge products. This demo makes the ranking mechanics visible so stakeholders can evaluate retrieval quality, resilience, and business relevance together.