Installation

Install Local GPT and its dependencies to start using AI in your Obsidian vault

Prerequisites

Before installing Local GPT, ensure you have:

• Obsidian version 0.15.0 or higher
• An AI provider configured (Ollama, OpenAI, or OpenAI-compatible API)

Local GPT requires the AI Providers plugin as a dependency. We’ll install this in the steps below.

Installation Steps

Step 1: Install AI Providers Plugin

The AI Providers plugin acts as a central hub for managing AI connections in Obsidian.

1. Open Community Plugins

   In Obsidian, open Settings → Community plugins → Browse

2. Search for AI Providers

   Search for AI Providers in the plugin browser

3. Install and Enable

   Click Install, then Enable the plugin

You can also install AI Providers directly from the Obsidian plugin store: AI Providers

Step 2: Install Local GPT Plugin

Obsidian Plugin Store (Recommended)

1. Open Community Plugins

   In Obsidian, go to Settings → Community plugins → Browse

2. Search for Local GPT

   Type Local GPT in the search bar

3. Install and Enable

   Click Install, then Enable to activate the plugin

Direct Link

Open Local GPT in the Obsidian plugin store

BRAT (Beta Testing)

If you want to test beta features or development versions, use BRAT:

1. Install BRAT

   Install the BRAT plugin from the community plugins

2. Add Local GPT

   In BRAT settings, click Add Beta Plugin and enter:

   pfrankov/obsidian-local-gpt
   
3. Enable the Plugin

   Enable Local GPT in Settings → Community plugins

BRAT installs pre-release versions that may contain bugs. Use this method only if you want to test new features.

Step 3: Configure an AI Provider

With both plugins installed, you need to configure at least one AI provider.

Option 1: Ollama (Local, Recommended)

Ollama runs AI models completely offline on your machine.

1. Install Ollama

   Download and install Ollama from ollama.ai

2. Pull a Model

   Open a terminal and run:

   ollama pull gemma2
   

   Or try other models like llama3, mistral, qwen2, etc.

3. Configure in Obsidian

   1. Open Settings → AI Providers
   2. Click Add Provider → Ollama
   3. Set the URL to http://localhost:11434 (default)
   4. Select your model (e.g., gemma2:latest)
   5. Save the provider
4. Set as Main Provider

   In Settings → Local GPT, select your Ollama provider as the Main Provider

For vision support, pull a vision model:

ollama pull llava
# or
ollama pull bakllava

Then configure it as the Vision Provider in Local GPT settings.

Option 2: OpenAI API

Use OpenAI’s cloud models like GPT-4.

1. Get API Key

   Sign up at platform.openai.com and create an API key

2. Configure in Obsidian

   1. Open Settings → AI Providers
   2. Click Add Provider → OpenAI
   3. Enter your API key
   4. Select a model (e.g., gpt-4o-mini)
   5. Save the provider
3. Set as Main Provider

   In Settings → Local GPT, select your OpenAI provider as the Main Provider

Using OpenAI’s API sends your data to their servers and incurs usage costs.

Option 3: OpenAI-Compatible API

Connect to any OpenAI-compatible endpoint (LM Studio, LocalAI, etc.).

1. Start Your Server

   Ensure your OpenAI-compatible server is running (e.g., LM Studio, text-generation-webui with OpenAI extension)

2. Configure in Obsidian

   1. Open Settings → AI Providers
   2. Click Add Provider → OpenAI Compatible
   3. Enter the base URL (e.g., http://localhost:8080/v1)
   4. Add API key if required
   5. Select or enter the model name
   6. Save the provider
3. Set as Main Provider

   In Settings → Local GPT, select your provider as the Main Provider

Optional: Configure RAG (Enhanced Actions)

To enable context-aware responses using RAG, you need an embedding model:

1. Install an Embedding Model

   For Ollama users, pull an embedding model:

   # For English content
   ollama pull nomic-embed-text
   
   # For multilingual content
   ollama pull bge-m3
   
2. Configure Embedding Provider

   1. In Settings → AI Providers, add a new Ollama provider
   2. Select your embedding model (e.g., nomic-embed-text:latest)
   3. In Settings → Local GPT, set this as the Embedding Provider

RAG features work by analyzing your linked notes, backlinks, and PDFs to provide more contextual AI responses. Learn more in the RAG System guide.

Optional: Configure Hotkeys

Set up keyboard shortcuts for quick access to Local GPT features:

1. Open Hotkey Settings

   Go to Settings → Hotkeys

2. Search for Local GPT

   Type Local in the search bar to filter Local GPT commands

3. Assign Shortcuts

   • Local GPT: Show context menu - e.g., Cmd+M or Ctrl+M
   • Local GPT: Action Palette - e.g., Cmd+J or Ctrl+J

Verify Installation

To confirm everything is working:

1. Open a Note

   Create or open any note in your vault

2. Select Text

   Highlight a sentence or paragraph

3. Open Context Menu

   Right-click the selection (or use your hotkey) and look for Local GPT actions

4. Test an Action

   Try the Summarize action to verify your AI provider is working

If you see AI-generated content appear in your note, installation is complete!

Troubleshooting

No AI Providers showing up

Make sure the AI Providers plugin is installed and enabled. Restart Obsidian if needed.

Ollama connection failed

• Verify Ollama is running: ollama list in terminal
• Check the URL is http://localhost:11434
• Try pulling the model again: ollama pull gemma2

Actions not appearing in context menu

• Ensure Local GPT plugin is enabled
• Check that you have at least one Main Provider configured
• Try reloading Obsidian

Embedding provider errors

• Embedding providers are optional for basic functionality
• Ensure you’ve pulled an embedding model: ollama pull nomic-embed-text
• Verify the embedding provider is configured separately from the main provider

For more help, see the Troubleshooting guide.

Next Steps

Quickstart Guide

Learn how to use Local GPT effectively

AI Providers Setup

Advanced provider configuration options

Create Custom Actions

Build actions tailored to your workflow

Action Palette

Master the powerful Action Palette feature