I built a graph-memory layer on top of turbovec for local/constrained RAG — looking for feedback
A developer has created a new open-source tool called turbo-graph, which is a fork of the turbovec library. This tool enhances local Retrieval-Augmented Generation (RAG) by adding a graph memory index. The goal is to improve RAG performance in constrained environments by handling complex filtering and contextual constraints beyond simple vector similarity. AI
IMPACT Offers a new option for developers building constrained RAG applications locally.