Hi everyone,I’ve tested several RAG workflows, including the Agentic RAG workflow by Cole. Unfortunately, I’m still not happy with the quality of the results. I just came across this video: https://www.youtube.com/watch?v=yzMzRH1308Y – it explains that CAG might deliver better performance.Has anyone here worked with CAG? And if so, is there a CAG template available to try it out?Thanks!

About a week ago, Tavus released a amazing demo of Charlie, a chatbot, using its Conversational Video Interface. It is apparently built on three of its models: Sparrow, Raven, and Phoenix. I am keen on using these models to conduct real-time interviews in n8n but there is no native node support for this app. At the present time, the demo is a bit laggy, partly because it apparently processes the input in large chunks and latency becomes a problem as it formulates a response. As humans, we process audio input asysnchronously, that is, we process cognitable chunks of it as we hear them but I don't believe the state of the art in conversational video AI is there quite yet. Wait a few weeks. @Nate Herk would you consider doing a demo of how to use the http node to implement api calls to Tavus?

For a deeper discussion on optimizing this workflow, let's chat here. In this video, I break down how I built the most in-depth AI research system in n8n—capable of generating detailed research reports with all sources linked for just $0.50 per report! 🚀 You'll see: ✅ How the system gathers and analyzes data ✅ How it compiles a PDF research report ✅ Key optimizations to make it even more efficient Everything you need to set up this workflow is linked in the template that's attached to this post! Google Sheet Template Tavily PDF API OpenRouter