TUTORIAL: Obsidian + OLLAMA LOCAL AI
If you want AI-powered chat and text tools inside your Obsidian vault without sending your notes to the cloud, you can run everything locally using Ollama and the Local LLM Helper plugin. No API keys, no subscriptions, no vault content ever leaving your machine.
This guide walks through the full setup in four steps.
Step 1: Install Ollama Locally
Ollama is the engine that runs the LLM on your machine and exposes a local API for other apps to use.
Download it from the official site:
[ollama.com/download](https://ollama.com/download)
On Ubuntu/Linux, you can also install it directly from the terminal:
curl -fsSL https://ollama.com/install.sh | sh
Once installed, Ollama runs automatically in the background as a service, listening at http://localhost:11434.
Verify it's running:
systemctl status ollama
Next, pull a chat model and an embedding model:
ollama pull llama3.2
ollama pull nomic-embed-text
llama3.2 handles chat and text generation. nomic-embed-text is a lightweight model used for semantic search across your notes.
Confirm both downloaded correctly with:
ollama list
Pro Tip If you have an NVIDIA GPU, install the drivers before running the install script: Ollama auto-detects and uses the GPU when available.
Step 2: Install the Local LLM Helper Plugin
Open Obsidian, go to Settings → Community plugins, and make sure Restricted mode is turned off. Click Browse, search for "Local LLM Helper", then Install and Enable it.
This plugin lets Obsidian talk to any OpenAI-compatible local LLM server: including Ollama, LM Studio, and vLLM.
Step 3: Configure the Plugin
Open the plugin's settings page at Settings → Local LLM Helper.
Connection
Provider: Ollama
Chat/default server URL: http://localhost:11434
Embedding server URL: leave blank (falls back to the chat server)
Embedding API key: leave blank (not needed for local models)
Models
Chat model: llama3.2 (must exactly match your ollama list output)
Embedding model: nomic-embed-text (or click Browse to pick from installed models)
Temperature: 0.7 (default: lower for more precise answers, higher for more creative ones)
Max tokens: 1024 (raise this for longer responses)
Important The model names in the plugin must match exactly what shows up when you run ollama list in the terminal. If they don't match, the plugin can't reach the model.
Step 4: Ready to Start
Once your models are pulled and settings are saved, open the Command palette (Ctrl+P / Cmd+P), search for "Local LLM Helper", and choose an action: chat with your notes, summarize a page, or generate text.
Summary
Setting up a private AI assistant in Obsidian comes down to four steps:
Install Ollama and pull a chat model plus an embedding model.
Install the Local LLM Helper plugin from the community plugin browser.
Configure the connection and model settings to point at your local Ollama server.
Run it from the command palette and start chatting with your vault.
That's it: a private AI assistant running entirely on your own machine, with no API keys, no cloud calls, and no vault data leaving your computer.
