Show HN: Embedding classifier ~100x lighter than Jev, try it offline in-browser

Category: ai-ml

Tags: embeddings, intent-classification, browser-ml, local-first, semantic-search

Score: 5.8/10 (Innovation: 5, Technical: 5, Documentation: 7, Utility: 6)

A browser-based demo that performs local semantic intent classification using a small embedding model (MiniLM-L6-v2) and cosine similarity, eliminating the need for fine-tuning or API calls. It showcases a lightweight, privacy-preserving alternative to LLM-based intent detection, with potential for offline voice/command interfaces.

Target audience: frontend developers, backend developers, AI/ML engineers, hobbyists

Repository: https://lusob.github.io/embeddings-is-all-you-need/ · HTML · 1 stars

View on Hacker News