I've been working on a Python package to enable Agentic RAG with Vectara (disclosure: I work there). It uses LlamaIndex behind the scenes, and provides a quick and easy way to build Agentic RAG assistants.
I hope you find it useful, and please share any questions or feedback.