Copilot & Illama

The way to get AI‑powered search over your entire Obsidian vault using an LLM is to combine Ollama (local LLM runner) with an Obsidian plugin that builds semantic embeddings of your notes. Based on current sources, the strongest option is Copilot, and Ollama plugin.

Obsidian Copilot plugin (best “ChatGPT‑style” assistant for your notes) This plugin connects Obsidian to a local LLM (via Ollama). It gives you a sidebar chat that can answer questions about your vault, summarize notes, draft content, etc.

Models

ollama pull qwen2.5 ollama pull nomic-embed-text

What you get:

Chat with your notes Summaries, drafting, rewriting Local RAG (retrieval‑augmented generation)

🔧 Recommended local models (Ollama) Embeddings (for search) nomic-embed-text

Chat models llama3.2 (fast, general‑purpose) qwen2.5 (excellent reasoning) gemma2 (good long‑context)

🧩 Example: Ask your vault a question Once Smart Connections is installed and indexed:

Prompt: What are the main themes across my research notes on LLMs?

Result: The plugin retrieves relevant note chunks, feeds them to the LLM, and returns a grounded answer with links to the exact notes.