This interactive demo shows how vector embeddings can be used to search text semantically rather than by exact keywords. This lies at the heart of how memory works in modern AI applications!
When you type a query below, the system converts your text into a numerical vector, compares it against precomputed embeddings of Alice's Adventures in Wonderland, and retrieves the most similar passages.

Try searching concepts like “rabbit hole”, “tea party”, or “Queen of Hearts” to see how semantically close passages are found.