Search saved video descriptions and text fetched from resources directly linked in those descriptions. No video media, captions, or transcripts are collected or indexed. No LLM API, embeddings, paid ...