Hi everyone, I’m a beginner software developer and...
# agents
f
Hi everyone, I’m a beginner software developer and recently started working with n8n workflows. I’ve now transitioned to using LiveKit for voice-based agents in a few portfolio projects. In n8n, I was used to building workflows via drag-and-drop and configurations in the browser. Now, as I’m considering self-hosting with LiveKit, I’m wondering: Is it possible to integrate RAG (retrieval-augmented generation) and custom tools? I got a bit confused by the documentation, which is why I’m reaching out to the community for some clarification. Any feedback would be greatly appreciated! Thanks in advance!
r
Here is a working RAG example you can start with: https://github.com/livekit-examples/python-agents-examples/tree/main/rag Also, we have some features being released next week that should help your situation. Stay tuned to our DevDay next week https://www.linkedin.com/feed/update/urn:li:activity:7394457283356876801/
f
So to better understand it!: Right now there is no capability of RAG in the LiveKit agents? and will be rolled out soon? And what about custom tools? How can we define it in the LiveKit agents?
r
Yes, you can do rag in LiveKit agents today. That is what the example I shared does.
Custom tools are also easy to use in LiveKit agents
MCP is also a common way to do it too
This basic_agent.py has an example of custom tool to get weather for your current location
I think you may gain a lot to go through this building voice agents guide: https://livekit.worksh.app/tutorials/livekit-voice-agent/introduction It will take thought getting started and continue with custom tools and MCP
And (as a teaser) next week we will have some other great features being released that will make all this even better.
f
So we're using RAG as a external tool. right? Not as a built in feature, that's what I'm asking!!
The chunking, embeddings etc, this is all handled outside, right?
r
Yes, to do RAG you will need something like a vector DB somewhere (external) and build the index in the vector store that can be queried by your agent.
✅ 1
f
Ty, understood!
👍🏼 1
Much helpful