memery

Why we included this project

Anyone who has a folder full of images knows the pain of scrolling through hundreds of thumbnails to find one specific picture. Memery lets you type a plain-language description, like "a line drawing of a woman facing left," and get relevant matches back from your local collection. It builds on OpenAI's CLIP model to embed both your images and your query into a shared space, then ranks the folder accordingly. You can browse results in a browser GUI or pipe them through command-line tools, and the library also works directly in Jupyter notebooks. That makes it useful for quick dataset exploration and for personal photo archives, and teams with unlabeled image corpora get a lightweight way to add semantic search without standing up a full retrieval stack.