"Word2vec for anything" is where we want to get to. Right now we only support images and text, but you can see the other data types on our roadmap at https://www.basilica.ai/available-embeddings/ .