New: Cookbooks and AI ExplanationsStep-by-Step recipes to solve problems connected to Roadmaps and Cheat Sheets. Need more details? Use AI buttons for structured and simple explanations with concrete examples throughout the whole platform.Take a look
I know I have the photo somewhere, but none of my files or folders are named anything close to what's in it.
What you'll have at the end
A search box that returns the right photo for a typed phrase like 'a dog on a beach', no file tags needed
You need
A small folder of your own photo files, whatever ordinary camera names they already carry, plus a way to run CLIP so it can turn both a photo and a typed phrase into a vector, whether through a hosted call or an installed model.
Not covered
Keeping the saved vectors in sync as photos are added or deleted later, and scaling past a personal-sized folder; this covers a first working search over a fixed set of photos.
Verify your first embedding works
go there first if you aren't yet sure a plain embedding call is returning usable vectors at all
Choose cosine similarity or dot product for your vectors
go there for the deeper case behind picking cosine similarity here instead of a raw dot product
Speed up nearest neighbor search with a FAISS index
go there once your own photo folder grows past what scoring every single photo on each search comfortably handles
Checked 18 Aug 2026
Part of the Embeddings cookbook