Vector embeddings are key to ML, and here we describe how to use OpenCV, OpenAI CLIP and pgvector to generate vectors and use them to perform image recognition on a corpus of photos.