About this tool
Paste two embedding vectors to compute cosine similarity, Euclidean distance, and dot product — a quick sanity check for RAG and semantic-search work. 100% client-side.
Frequently asked questions
What is cosine similarity and why use it for embeddings?
Cosine similarity measures the angle between two vectors, ignoring their magnitude — it ranges from -1 (opposite) to 1 (identical direction). It's the standard metric for embeddings because most embedding models encode meaning in the *direction* of the vector, not its length, so cosine similarity captures semantic closeness better than raw distance.
Can I compare embeddings from different models?
Generally, no — this tool will happily compute a number if the vectors are the same length, but embeddings from different models live in different, incompatible vector spaces (even if dimensions happen to match). A meaningful similarity score requires both vectors to come from the exact same embedding model.