buraksenyurt/rust-llm-wiki

A personal prototype of Andrej Karpathy’s “LLM Wiki” idea implemented with Obsidian and Microsoft Copilot CLI, built around my Rust blog posts. Purpose: convert raw articles into a small LLM-backed wiki, summarize semantic relations, and query the archive via Copilot commands — effectively a lightweight RAG-like workflow without embeddings, vector DBs, or costly retrieval pipelines. Who uses it: individual creators or learners who want searchable, summarized notes and code examples (e.g., learning a new language); possibly useful for small teams but scalability and trust need evaluation. Key features: - Folder layout: /raw, /templates, /wiki plus COPILOT.md instructions. - Copilot CLI \init to generate instruction file and update wiki from raw content. - Natural-language prompts to refresh wiki, lint formatting/links, detect orphan docs, and create/update cross-links. - Obsidian Graph visualization of relations. - Session memory speeds up responses. Limitations: requires an AI agent (Copilot/Claude), output quality depends on raw content, and results must be verified.

AI-related keywords found in this repo's description, topics, or README summary — grouped by category. Each badge links to the corresponding ranking detail page.