Memery

Memery is an open source tool that lets users search a collection of images using ordinary language instead of folders and filenames. It was created by the developer who writes as deepfates, who describes years of frustration finding a particular meme, photo or screenshot on a hard drive and built Memery to remove that friction. Under the hood it uses OpenAI's CLIP transformer to match text queries to image content. Queries can name concrete things such as a bird or a wizard, abstract concepts such as love or a graph, or text that appears inside an image. Memery indexes a folder of images, stores the encodings so later searches are quick, and returns the best matches for a query. It can be used in several ways. It works as a Python library, where a single function call takes a folder and a query and returns matching file paths. It also includes a browser based graphical interface and a command line interface. The project is hosted on GitHub, where the author invites others to test it, file issues and send pull requests. Memery suits developers, meme collectors and anyone with a large local image library who wants semantic search that runs on their own files. The listing points to the author's personal blog post from July 2021 introducing the project rather than to a dedicated product site. No pricing is published, and the project is presented as open source software.