So I've been going down a rabbit hole lately and I can't find many people actually talking about this specific use case.
everyone here runs local LLMs for coding, chat, maybe some creative writing. cool. But what about using it as a proper personal knowledge base? like, dump your own notes, PDFs, random docs into it and actually query your own life privately, every day.
I tried looking into this seriously and hit a wall. Most resources either assume you're a developer building something, or they're 2 years old and recommend tools that have completely changed since.
So genuinely asking, is anyone here actually doing this day to day? Not as an experiment, but as a real workflow?
Things I keep running into that I can't figure out:
- What model are you running for this? RAG on consumer hardware seems finicky depending on quant
- Do you actually trust the retrieval or do you double check everything because hallucinations?
- LlamaIndex vs Ollama vs whatever else has anything actually made this less painful recently?
- Context length, how do you handle it when your personal docs start piling up?
Not looking for a tutorial or a GitHub repo. Just want to hear from someone who's made this work without it becoming a part time job to maintain.
[link] [comments]