Show HN: A 34 MB search engine that can't make a network call (www.korvo.xyz)

A groundbreaking new search engine, Korvo Embed, has been introduced, featuring a compact 34 MB model that operates entirely offline on your machine. This system is designed to enhance information retrieval by focusing on intent rather than keywords, meaning users can get relevant search results even if their queries don’t match the exact words on the page. Notably, this model is configured with 384 dimensions per chunk and can process 512 tokens at a time, all without making network calls, ensuring user privacy. The significance of Korvo Embed lies in its strong emphasis on privacy and local data processing, as it stores all indexing and search functions directly on the device without third-party involvement. With a built-in cryptographic verification process for its model and a design that prevents any document exfiltration, it offers businesses a reliable local-first software solution that complies with stringent security reviews. Future updates promise enhanced capabilities like fast indexing, reranking for improved result quality, and additional language support, solidifying its potential impact on the AI/ML community and industries requiring high data security.