LLMWiki

LLM Wiki is a persistent knowledge workflow built inside an Obsidian vault.

Instead of answering questions from raw files from scratch every time, it maintains a durable intermediate layer of understanding in wiki/.

• raw/ stores the source material
• wiki/ stores the maintained knowledge
• the agent reads, synthesizes, links, and updates
• the wiki becomes more useful over time

Why This Is Different From Standard RAG

Traditional RAG usually works like this:

• store raw files
• retrieve chunks at question time
• generate an answer
• repeat the same reconstruction later

LLM Wiki works differently:

• capture source material into raw/
• ingest it into a structured wiki
• improve the wiki over time
• answer future questions from that maintained knowledge layer

The main shift is this: standard RAG is often retrieval-first, while LLM Wiki is maintenance-first.

That means the system is not only trying to answer the current question. It is trying to build durable understanding that can be reused later.

Why It Feels Like A Game Changer

This approach matters because it turns AI from a one-time responder into an ongoing knowledge maintainer.

That changes the experience in a few important ways: