{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/a5cb417115684a3e82a5fd3f20266489\" frameborder=\"0\" width=\"830\" height=\"622\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":622,"width":830,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":622,"thumbnail_width":830,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/a5cb417115684a3e82a5fd3f20266489-00001.gif","duration":292.2218,"title":"Natural Language Search in the Browser","description":"https://jasonjmcghee.github.io/portable-hnsw/\n\nIn this video, I demonstrate a solution I developed over the weekend that allows for natural language search in the browser without the need for a vector database. I show how you can search for specific content based on fuzzy semantic meaning, using a custom index algorithm and transformers JS. The solution offers flexibility and scalability, allowing for large indexes and efficient search queries. No action is requested from the viewers, but I invite you to explore this new approach to browser-based search."}