LLM-WIKI

LLM-WIKI is an Obsidian-style knowledge graph and retrieval chatbot for computational biology lecture material. It converts lecture notes, source material, concepts, entities, methods, equations, and comparisons into a structured Markdown wiki that can be searched by an LLM-backed chat interface.

The project has three main parts:

• llm-wiki/: the Markdown knowledge graph, Obsidian vault, graph-repair scripts, retrieval files, and wiki quality tools.
• backend/: a FastAPI RAG server that searches the wiki and exposes OpenAI-compatible chat endpoints.
• frontend/: a Next.js chat UI adapted from Chatbot UI with a no-auth LLM Wiki chat page.

Knowledge Graph Preview

Click the graph image to open the full-size SVG and zoom into the LLM-WIKI concept network.

What This Project Does

This repo is designed to make computational biology lecture knowledge easier to browse, connect, and query.

• Obsidian graph exploration of computational biology concepts.
• Canonical lecture, concept, method, entity, equation, comparison, and topic notes.
• Wiki scoring, duplicate detection, orphan repair, graph indexing, and link validation scripts.
• A FastAPI backend that retrieves relevant wiki chunks.
• A Next.js frontend chat page that asks questions against the wiki.
• OpenAI-compatible /v1/chat/completions API behavior for easier integration with chat clients.

How The Frontend And Backend Connect

The frontend runs in the browser on port 3000.