Oh word? What do you all like about N8N and Pinecone?

Yeah n8n is nice! I started using it earlier this year. I mostly use it as a 24/7 scheduler. I know I could do it other ways like a cron job or something but I like n8n. I self host mine on a raspberry pi so its always on and running. lets my automations run uninterrupted without much limitation. But I mostly just have experiments that pipe out to a custom web server I built right now. I have tried some full workflows in n8n itself though. I have played around with generating n8n workflows with LLMs thats pretty incredible. And recently the MCP feature in n8n is my current interest lol. Its pretty crazy because I am just using it as a universal MCP server its super easy to add new/remove services from there. I want to try connecting ChatGPT's new MCP to an n8n instance MCP and see how crazy that can get 👀😎 I havent done it yet though I want to make sure its all done securely so I gotta get around to that. lmao anyways. What about you?

Sounds pretty cool! What use cases are you solving for @powerful-air-75070 ? Do you have any examples you can share?

you self host yours on a RASPBERRY PI??!?!

Yeah for sure ! And I have had 0 problems in like 6 months at least. Its perfect for it. You can lock the pi down and then put it on a tailnet using https://tailscale.com/ and then you can have secure zero trust api access through your tailnet. Its all free too. open source and free tier

which workflows have you used Pinecone in? and what are you using a vdb for?

@polite-carpenter-85994 Yeah so how I stumbled on Pinecone was separate from my n8n work. Long story short I was trying to replicate ChatGPT memory functionality locally. I was using Qdrant. I have a automation pipeline that I wanted to try converting to a SaaS. Thats how I discovered Pinecone because I need multi user support for a vector database in production. Essentially I have a content generation pipeline mostly for Wordpress. Part of the user journey includes importing context from integrations such as Notion, Wordpress, Github, Docs, various context sources. I am going to chunk and embed these when the user wants to include them in their "Context Vault". So when the user's pipeline is fully executed it will use the users context embeddings and RAG to generate their content. This gives your generated content the anchored "rooted in your actual work" kind of feel. It works pretty good for me.

Did Qdrant not support multi-users or just the number of namespaces that you need?

i haven't really gotten into n8n. but i'd love to learn more

hmm well to be honest I think I just started gravitating towards Pinecone because it reminds me of Supabase and I have had such a positive experience with Supabase.

That's awesome, what a cool flipping n8n flow! Gonna wanna stare at it a bit more to learn and understand. Maybe I should buy a rasberry pi... hahha

@brainy-leather-20096 Yeah not a lot of recent work with n8n + Pinecone specifically right now. But over the weekend I did make myself a small WordPress plugin for Pinecone so that way whenever I publish a new post that post automatically gets chunked and embedded in Pinecone! Then Im using those embeddings in a RAG chatbot and in an Algolia like site wide search. Im writing tutorials on all of it this week. So thats some of what Im working on right now but cool/fun project over the weekend! 😎

Ok love it. I’ve wondered what kind of support Wordpress has these days for AI tooling and plugins. It’s been a decade plus since I did anything WP related. Definitely reach out if you run into any snags and share some links when you’re ready. We’ll find a place to share them more broadly too.

Absolutely will do and thank you I appreciate it!