hey all - thanks for the welcome! I am currently building a local and private research knowledge base so I can fire any tweet links, youtube videos, web sites or github repos at it easily and have it index and cross reference itself, then be sensibly navigated or queried. eventually this might seek out information itself, and I plan to also link it to my phone. tools are obsidian, lots of python code from github copilot, and a hermes agent over top as a second stage. Previously I tried to use onenote or raindrop as stores of information, and was frustrated with the lack of twitter api for extracting bookmarks - so it looks like obsidian on desktop and mobile is the way to reduce friction in gathering, and hermes agent looks like a great layer to automate or semi automate research and presentation, especially with playwright for web access. I'm currently doing an initial indexing and enrichment pass on 50k+ links i have collected, and had to build a robust dashboard to run it on all my cpu cores with thermal tuning and quality control options (thanks copilot). That will take at least a day! Very interested to see all the personal project angles here :)