As a novice only a few short months ago, I needed a way to impement embeddings which was easy and relatively inexpensive. WCS came through on both of those for me. In addition, as I began to gain more experience, I was able to appreciate more the vast array of embed and retrieval options WCS had to offer. Now, as a grizzled veteran now deploying my 3rd RAG system, I am extremely happy with the decision I made early on to go with WCS. Integrating into my code as very easy. Their support is great and timely, and they appear to spend a good amount of effort continuously improving an already superior system. In RAG, your answers are only ever going to be as good as the documents your model is given to analyze, and I have to say that WCS cosine similarity searches have consistently given me back the best documents for the best answers.

* Simple to get a local dev env working with docker * Their new v4 python sdk makes things way easier to work with * Built-in hybrid search with the options to tailor it to exactly what you need is a huge plus * Reliable, been using it in production for 6 months without a hiccup

It was very easy to get started with weaviate, the pricing is simple and they take care of importing models from huggingface for us and provisioning the db instances so we can concentrate on building our product. We were able to quickly integrate weaviate with our existing stack using the python libraries. The documentation is good and we have been impressed by the support we received for our more unusual use cases. There is an active slack community and we've met several people from the team and they've all been very helpful. What's more it's all open source so we feel safe choosing Weaviate for a cricical part of our tech stack.