It's like SQLITE but for embeddings, introducing VDBLITE

Hey everyone, there’s an unmet need for rapid prototyping applications that need vector search (search for embeddings). Sure there are highly scalable things like Milvus and FAISS but they lack features or require servers. I wanted something that was local and fast enough.

So I made VDBLITE. It’s still in alpha, but it’s functional and useful. It allows you to quickly search in memory or save to file. Furthermore, it allows for arbitrary contents of the elements - you define them however you like.

Lots of opportunity for optimization here, this is just the beginning. But it’s ready in pypi so you can install it with just pip install vdblite and you’re off to the races.