I’ve been spending a bit of my evenings in the last six months playing around with documents I have collected over the last couple of decades of work and reading up on how to build each step right, especially as someone who is new to actually coding and building. Really have learned to love the idea of how weaviate works and coudln’t get my own implementation going but I found DIFY and I was pretty happy with it. Considering the wealth of documents, I’ve divided up my docs into similar categories and uploaded them as knowledges into DIFY. I’ve had some good results on some collections but weak results on others. I also realized that DIFY doesn’t really utilize all of the things that weaviate has to offer.
I’ve love to get a sense of what you guys think of DIFY, it’s shortcomings as it relates to weaviate and if there are other full stack repos out there that use weaviate.
I should say, I realize that I need to try Verba again. I use nomic for embeddings, and various models from ollama locals to claude for the LLM. I’ve only really had satisfying results with GPT4 and the new Claude.
