Loading...

KJ5ILQ returning. You can. That would be an option. I tend to be a little more granular than that. I will probably set up some tasks under an AI agent workflow that probably through an MCP. Oh, you know what? Obsidian, I don't think, released the MCP yet. They will very soon. So I'll probably have to go through my pie to sync all this. Anyway, that's fine. My idea is this actually. I'm going to have a data dump note inside of the obsidian vault to where everything just gets dumped. Everything. But then I'll run an agent behind the scenes that will organize it and create links, notes, tags, and keep everything in the markdown format after the fact. So really the way that I'm picturing it is through, I'm going to use Rube. R-U-B-E. If you've never heard of it, it's kind of like N8N. It's a connector. It's a connector. And it's very simple. Very, very simple. I actually like it. So anyhow, I have Rube connected to my chat GPT. I use that as an example. I have it connected to more than just chat GPT. So anyhow, it's connected to that. So then I can connect my Raspberry Pi in the obsidian vault through the Rube, which will then interact with the LLM. So I can still set up tasks or I can set up a bot, just a separate bot that does nothing but take that drop-in file. And the first thing that it does is take it and converts it over to markdown. And then I have another agent in the path line that would then take that markdown and separate it into, let's say, tags. And then another one that would separate links and backlink itself. That way you have all the functionality of obsidian being used by AI so I don't have to go in there and then physically change it into a markdown file or physically make links from one note to another. That's kind of my idea. Now, like I said, I've just been researching the idea of it for a day or two, so none of it's in practice yet. I think it will work.